Systems and methods for classifying objects in digital images captured using mobile devices

Description

FIELD OF INVENTION

The present invention relates to mobile image capture and image processing, and more particularly to capturing and processing digital images using a mobile device, and classifying objects detected in such digital images.

BACKGROUND OF THE INVENTION

Digital images having depicted therein an object inclusive of documents such as a letter, a check, a bill, an invoice, etc. have conventionally been captured and processed using a scanner or multifunction peripheral coupled to a computer workstation such as a laptop or desktop computer. Methods and systems capable of performing such capture and processing are well known in the art and well adapted to the tasks for which they are employed.

However, in an era where day-to-day activities, computing, and business are increasingly performed using mobile devices, it would be greatly beneficial to provide analogous document capture and processing systems and methods for deployment and use on mobile platforms, such as smart phones, digital cameras, tablet computers, etc.

Traditionally, digital images have been a valuable resource of data for a nearly infinite variety of applications. In a business context, digital images have been extensively utilized for communicating and processing information, typically represented in documents and/or associated image data (such as a digital image of a vehicle associated with a digital image of an insurance claim, vehicle registration, bill of sale, etc.). Increasingly powerful mobile devices offer opportunities to expand digital image processing into the mobile arena and provide improved capability to capture and process digital image data in real-time using mobile technology.

Conventional data extraction methods for use in existing mobile and non-mobile devices rely on object templates, typically generated and/or curated by expert users, to provide information to an extraction engine instructing the engine where to locate information for extraction. In the particular case of documents, the conventional extraction technology is provided the location of one or more (typically rectangular) regions of a document, instructed to perform optical character recognition (OCR) on the region(s), and then output the determined characters to another process or file.

While the conventional extraction engines and methods are capable of reliably extracting information from objects for which a template has been previously defined, it is not possible to dynamically extract information from objects about which no template exists. This is an undesirable limitation that restricts users from using powerful extraction technology on an increasingly diverse array of documents encountered in the modern world.

Furthermore, conventional extraction engines require extensive input from expert curators to define templates and maintain template definitions as object classes evolve. The performance of template-based extraction is thus a direct function of the curators' ability to properly define templates and the curators' determination of which information is “worth” extracting. Therefore, expert curators serve as an undesirable bottleneck on the robustness of data extraction in terms of extraction accuracy and precision, as well as the scope of objects from which data may be extracted.

Further still, conventional extraction methods rely primarily or exclusively on OCR techniques to extract text characters from image data. The OCR engine is forced to make estimates regarding the identity of text characters, which inevitably leads to erroneous reporting of characters when image quality is poor, when characters do not match a predefined set of “known” characters, when an apparent character appears ambiguous such that the OCR engine cannot reliably distinguish between multiple candidate characters (e.g. a period “.“versus a comma”.” or a letter “l” versus a numeral “1”). Expert curators can mitigate these problems by urging the OCR engine toward the correct decision when certain known patterns of characters are expected, but even this mitigation is limited in scope and errors ultimately require undesirable end-user interaction.

Therefore, it would be highly beneficial to provide new method, system and/or computer program product technology for extracting information from digital image data using mobile devices. It would be further beneficial to enable extraction of information without relying on templates, thus circumventing the need for expert curators and enabling users to dynamically generate and modify extraction models to extract data from diverse and mutable classes of objects. It would be still further beneficial to provide methods for extracting data without relying on OCR techniques to overcome the limitations of predefined character classes, poor image quality, and reduce or remove the need for user correction of OCR mistakes.

SUMMARY OF THE INVENTION

In one embodiment, a method includes: receiving or capturing digital image data; detecting an object depicted in a digital image represented by the digital image data; processing the digital image to improve an image quality of at least the object; determining an object classification based on the processed digital image; building or determining an extraction model based at least in part on the object classification; and extracting, using the extraction model, information from one or more regions of interest within the digital image.

In another embodiment, a mobile device is configured to execute logic. The mobile device includes logic specifically configured to cause the mobile device, upon execution of the logic, to: receive or capture digital image data; detect an object depicted in a digital image represented by the digital image data; process the digital image to improve an image quality of at least the object; determine an object classification based on the processed digital image; build or determine an extraction model based at least in part on the object classification; and extract, using the extraction model, information from one or more regions of interest within the object.

In yet another embodiment, a computer program product includes a computer readable storage medium having embodied thereon executable instructions configured to cause a mobile device, upon execution of the instructions, to: receive or capture digital image data; detect an object depicted in a digital image represented by the digital image data; process the digital image to improve an image quality of at least the object; determine an object classification based on the processed digital image; build or determine an extraction model based at least in part on the object classification; and extract, using the extraction model, information from one or more regions of interest within the object.

Other aspects and features of the presently disclosed inventive concepts will become apparent from the following detailed descriptions, which should be understood to be illustrative in nature and not limiting on the instant disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates a network architecture, in accordance with one embodiment.

FIG. 2 shows a representative hardware environment that may be associated with the servers and/or clients of FIG. 1, in accordance with one embodiment.

FIG. 3A depicts a digital image of an object, according to one embodiment

FIG. 3B depicts a schematic representation of the digital image shown in FIG. 3A divided into a plurality of sections for generating a first representation of the digital image, according to one embodiment.

FIG. 3C is depicts a first representation of the digital image shown in FIG. 3A, the first representation being characterized by a reduced resolution relative to the resolution of the digital image.

FIG. 4A is a schematic representation of a plurality of subregions depicted in a digital image of a document, according to one embodiment.

FIG. 4B is a masked representation of the digital image shown in FIG. 4A, according to one embodiment.

FIG. 4C is a masked representation of the digital image shown in FIG. 4A, according to one embodiment.

FIG. 4D is a masked representation of the digital image shown in FIG. 4A, according to one embodiment.

FIG. 5 is a flowchart of a method, according to one embodiment.

FIG. 6 is a flowchart of a method, according to one embodiment.

FIG. 7 is a flowchart of a method, according to one embodiment.

FIG. 8 is a flowchart of a method, according to one embodiment.

DETAILED DESCRIPTION

The following description is made for the purpose of illustrating the general principles of the present invention and is not meant to limit the inventive concepts claimed herein. Further, particular features described herein can be used in combination with other described features in each of the various possible combinations and permutations.

Unless otherwise specifically defined herein, all terms are to be given their broadest possible interpretation including meanings implied from the specification as well as meanings understood by those skilled in the art and/or as defined in dictionaries, treatises, etc.

It must also be noted that, as used in the specification and the appended claims, the singular forms “a,” “an” and “the” include plural referents unless otherwise specified.

The present application refers to image processing of images (e.g. pictures, figures, graphical schematics, single frames of movies, videos, films, clips, etc.) captured by cameras, especially cameras of mobile devices. As understood herein, a mobile device is any device capable of receiving data without having power supplied via a physical connection (e.g. wire, cord, cable, etc.) and capable of receiving data without a physical data connection (e.g. wire, cord, cable, etc.). Mobile devices within the scope of the present disclosures include exemplary devices such as a mobile telephone, smartphone, tablet, personal digital assistant, iPod®, iPad®, BLACKBERRY® device, etc.

However, as it will become apparent from the descriptions of various functionalities, the presently disclosed mobile image processing algorithms can be applied, sometimes with certain modifications, to images coming from scanners and multifunction peripherals (MFPs). Similarly, images processed using the presently disclosed processing algorithms may be further processed using conventional scanner processing algorithms, in some approaches.

Of course, the various embodiments set forth herein may be implemented utilizing hardware, software, or any desired combination thereof. For that matter, any type of logic may be utilized which is capable of implementing the various functionality set forth herein.

One benefit of using a mobile device is that with a data plan, image processing and information processing based on captured images can be done in a much more convenient, streamlined and integrated way than previous methods that relied on presence of a scanner. However, the use of mobile devices as document(s) capture and/or processing devices has heretofore been considered unfeasible for a variety of reasons.

In one approach, an image may be captured by a camera of a mobile device. The term “camera” should be broadly interpreted to include any type of device capable of capturing an image of a physical object external to the device, such as a piece of paper. The term “camera” does not encompass a peripheral scanner or multifunction device. Any type of camera may be used. Preferred embodiments may use cameras having a higher resolution, e.g. 8 MP or more, ideally 12 MP or more. The image may be captured in color, grayscale, black and white, or with any other known optical effect. The term “image” as referred to herein is meant to encompass any type of data corresponding to the output of the camera, including raw data, processed data, etc.

As will be appreciated by one skilled in the art, aspects of the present invention may be embodied as a system, method or computer program product. Accordingly, aspects of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, micro-code, etc.) or an embodiment combining software and hardware aspects that may all generally be referred to herein as “logic,” “circuit,” “module” or “system.” Furthermore, aspects of the present invention may take the form of a computer program product embodied in one or more computer readable medium(s) having computer readable program code embodied thereon.

Any combination of one or more computer readable medium(s) may be utilized. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, processor, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband, as part of a carrier wave, an electrical connection having one or more wires, an optical fiber, etc. Such a propagated signal may take any of a variety of forms, including, but not limited to, electromagnetic, optical, or any suitable combination thereof. A computer readable signal medium may be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like and conventional procedural programming languages, such as the “C” programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider).

Aspects of the present invention are described below with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer readable medium that can direct a computer, other programmable data processing apparatus, or other devices to function in a particular manner, such that the instructions stored in the computer readable medium produce an article of manufacture including instructions which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

FIG. 1 illustrates an architecture 100, in accordance with one embodiment. As shown in FIG. 1, a plurality of remote networks 102 are provided including a first remote network 104 and a second remote network 106. A gateway 101 may be coupled between the remote networks 102 and a proximate network 108. In the context of the present architecture 100, the networks 104, 106 may each take any form including, but not limited to a LAN, a WAN such as the Internet, public switched telephone network (PSTN), internal telephone network, etc.

In use, the gateway 101 serves as an entrance point from the remote networks 102 to the proximate network 108. As such, the gateway 101 may function as a router, which is capable of directing a given packet of data that arrives at the gateway 101, and a switch, which furnishes the actual path in and out of the gateway 101 for a given packet.

Further included is at least one data server 114 coupled to the proximate network 108, and which is accessible from the remote networks 102 via the gateway 101. It should be noted that the data server(s) 114 may include any type of computing device/groupware. Coupled to each data server 114 is a plurality of user devices 116. Such user devices 116 may include a desktop computer, lap-top computer, hand-held computer, printer or any other type of logic. It should be noted that a user device 111 may also be directly coupled to any of the networks, in one embodiment.

A peripheral 120 or series of peripherals 120, e.g., facsimile machines, printers, networked and/or local storage units or systems, etc., may be coupled to one or more of the networks 104, 106, 108. It should be noted that databases and/or additional components may be utilized with, or integrated into, any type of network element coupled to the networks 104, 106, 108. In the context of the present description, a network element may refer to any component of a network.

According to some approaches, methods and systems described herein may be implemented with and/or on virtual systems and/or systems which emulate one or more other systems, such as a UNIX system which emulates an IBM z/OS environment, a UNIX system which virtually hosts a MICROSOFT WINDOWS environment, a MICROSOFT WINDOWS system which emulates an IBM z/OS environment, etc. This virtualization and/or emulation may be enhanced through the use of VMWARE software, in some embodiments.

In more approaches, one or more networks 104, 106, 108, may represent a cluster of systems commonly referred to as a “cloud.” In cloud computing, shared resources, such as processing power, peripherals, software, data, servers, etc., are provided to any system in the cloud in an on-demand relationship, thereby allowing access and distribution of services across many computing systems. Cloud computing typically involves an Internet connection between the systems operating in the cloud, but other techniques of connecting the systems may also be used.

FIG. 2 shows a representative hardware environment associated with a user device 116 and/or server 114 of FIG. 1, in accordance with one embodiment. Such figure illustrates a typical hardware configuration of a workstation having a central processing unit 210, such as a microprocessor, and a number of other units interconnected via a system bus 212.

The workstation shown in FIG. 2 includes a Random Access Memory (RAM) 214, Read Only Memory (ROM) 216, an I/O adapter 218 for connecting peripheral devices such as disk storage units 220 to the bus 212, a user interface adapter 222 for connecting a keyboard 224, a mouse 226, a speaker 228, a microphone 232, and/or other user interface devices such as a touch screen and a digital camera (not shown) to the bus 212, communication adapter 234 for connecting the workstation to a communication network 235 (e.g., a data processing network) and a display adapter 236 for connecting the bus 212 to a display device 238.

The workstation may have resident thereon an operating system such as the Microsoft Windows® Operating System (OS), a MAC OS, a UNIX OS, etc. It will be appreciated that a preferred embodiment may also be implemented on platforms and operating systems other than those mentioned. A preferred embodiment may be written using JAVA, XML, C, and/or C++ language, or other programming languages, along with an object oriented programming methodology. Object oriented programming (OOP), which has become increasingly used to develop complex applications, may be used.

An application may be installed on the mobile device, e.g., stored in a nonvolatile memory of the device. In one approach, the application includes instructions to perform processing of an image on the mobile device. In another approach, the application includes instructions to send the image to a remote server such as a network server. In yet another approach, the application may include instructions to decide whether to perform some or all processing on the mobile device and/or send the image to the remote site.

In various embodiments, the presently disclosed methods, systems and/or computer program products may utilize and/or include any of the functionalities disclosed in related U.S. patent application Ser. No. 13/740,123, filed Jan. 11, 2013. For example, digital images suitable for processing according to the presently disclosed algorithms may be subjected to any image processing operations disclosed in the aforementioned Patent Application, such as page detection, rectangularization, detection of uneven illumination, illumination normalization, resolution estimation, blur detection, etc.

In more approaches, the presently disclosed methods, systems, and/or computer program products may be utilized with, implemented in, and/or include one or more user interfaces configured to facilitate performing any functionality disclosed herein and/or in the aforementioned related Patent Application, such as an image processing mobile application, a case management application, and/or a classification application, in multiple embodiments.

In still more approaches, the presently disclosed systems, methods and/or computer program products may be advantageously applied to one or more of the use methodologies and/or scenarios disclosed in the aforementioned related Patent Application, among others that would be appreciated by one having ordinary skill in the art upon reading these descriptions.

It will further be appreciated that embodiments presented herein may be provided in the form of a service deployed on behalf of a customer to offer service on demand.

Document Classification

In accordance with one inventive embodiment commensurate in scope with the present disclosures, as shown in FIG. 5, a method 500 is shown. The method 500 may be carried out in any desired environment, and may include embodiments and/or approaches described in relation to FIGS. 1-4D, among others. Of course, more or less operations than those shown in FIG. 5 may be performed in accordance method 500 as would be appreciated by one of ordinary skill in the art upon reading the present descriptions.

In operation 502, a digital image captured by a mobile device is received.

In one embodiment the digital image may be characterized by a native resolution. As understood herein, a “native resolution” may be an original, native resolution of the image as originally captured, but also may be a resolution of the digital image after performing some pre-classification processing such as any of the image processing operations described above and in copending U.S. patent application Ser. No. 13/740,123, filed Jan. 11, 2013, a virtual re-scan (VRS) processing as disclosed in related U.S. Pat. No. 6,370,277, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. In one embodiment, the native resolution is approximately 500 pixels by 600 pixels (i.e. a 500×600 digital image) for a digital image of a driver license subjected to processing by VRS before performing classification. Moreover, the digital image may be characterized as a color image in some approaches, and in still more approaches may be a cropped-color image, i.e. a color image depicting substantially only the object to be classified, and not depicting image background.

In operation 504, a first representation of the digital image is generated using a processor of the mobile device. The first representation may be characterized by a reduced resolution, in one approach. As understood herein, a “reduced resolution” may be any resolution less than the native resolution of the digital image, and more particularly any resolution suitable for subsequent analysis of the first representation according to the principles set forth herein.

In preferred embodiments, the reduced resolution is sufficiently low to minimize processing overhead and maximize computational efficiency and robustness of performing the algorithm on the respective mobile device, host device and/or server platform. For example, in one approach the first representation is characterized by a resolution of about 25 pixels by 25 pixels, which has been experimentally determined to be a particularly efficient and robust reduced resolution for processing of relatively small documents, such as business cards, driver licenses, receipts, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

Of course, in other embodiments, different resolutions may be employed without departing from the scope of the present disclosure. For example, classification of larger documents or objects may benefit from utilizing a higher resolution such as 50 pixels by 50 pixels, 100 pixels by 100 pixels, etc. to better represent the larger document or object for robust classification and maximum computational efficiency. The resolution utilized may or may not have the same number of pixels in each dimension. Moreover, the most desirable resolution for classifying various objects within a broad range of object classes may be determined experimentally according to a user's preferred balance between computational efficiency and classification robustness. In still more embodiments, any resolution may be employed, and preferably the resolution is characterized by comprising between 1 pixel and about 1000 pixels in a first dimension, and between 1 and about 1000 pixels in a second dimension.

One exemplary embodiment of inputs, outputs and/or results of a process flow for generating the first representation will now be presented with particular reference to FIGS. 3A-3C, which respectively depict: a digital image before being divided into sections (e.g. digital image 300 as shown in FIG. 3A); a digital image divided into sections (e.g. sections 304 as shown in FIG. 3B); and a first representation of the digital image (e.g. representation 310 as shown in FIG. 3C) characterized by a reduced resolution.

As shown in FIGS. 3A-3B, a digital image 300 captured by a mobile device may be divided into a plurality of sections 304. Each section may comprise a plurality of pixels 306, which may comprise a substantially rectangular grid of pixels such that the section has dimensions of ps_(x)horizontal pixels (ps_(x)=4 as shown in FIG. 3B) by ps_(y)vertical pixels (ps_(y)=4 as shown in FIG. 3B).

In one general embodiment, a method includes receiving or capturing a digital image using a mobile device; using a processor of the mobile device to: determine whether an object depicted in the digital image belongs to a particular object class among a plurality of object classes; determine one or more object features of the object based at least in part on the particular object class at least partially in response to determining the object belongs to the particular object class; build or select an extraction model based at least in part on the one or more determined object features; and extract data from the digital image using the extraction model.

In another general embodiment, a method includes receiving or capturing a digital image using a mobile device; and using a processor of the mobile device: determining whether an object depicted in the digital image belongs to a particular object class among a plurality of object classes; displaying the digital image on a display of the mobile device upon determining the object does not belong to any of the plurality of object classes; receiving user input via the display of the mobile device, the user input identifying one or more regions of interest in the object; building and/or selecting an extraction model based at least in part on the user input; and extracting data from the digital image based at least in part on the extraction model.

In yet another general embodiment, a computer program product includes a computer readable storage medium having program code embodied therewith, the program code readable/executable by a processor to: receive a digital image captured by a mobile device; determine whether an object depicted in the digital image belongs to a particular object class among a plurality of object classes; determine one or more object features of the object based at least in part on the particular object class upon determining the object belongs to the particular object class; build or select an extraction model based at least in part on the one or more object features; and extract data from the digital image based at least in part on the extraction model.

A first representation may be generated by dividing a digital image R (having a resolution of x_Rpixels by y_Rpixels) into S_xhorizontal sections and S_yvertical sections and thus may be characterized by a reduced resolution r of S_xpixels by S_ypixels. Thus, generating the first representation essentially includes generating a less-granular representation of the digital image.

For example, in one approach the digital image 300 is divided into S sections, each section 304 corresponding to one portion of an s-by-s grid 302. Generating the first representation involves generating a s-pixel-by-s-pixel first representation 310, where each pixel 312 in the first representation 310 corresponds to one of the S sections 304 of the digital image, and wherein each pixel 312 is located in a position of the first representation 310 corresponding to the location of the corresponding section 304 in the digital image, i.e. the upper-leftmost pixel 312 in the first representation corresponds to the upper-leftmost section 304 in the digital image, etc.

Of course, other reduced resolutions may be employed for the first representation, ideally but not necessarily according to limitations and/or features of a mobile device, host device, and or server platform being utilized to carry out the processing, the characteristics of the digital image (resolution, illumination, presence of blur, etc.) and/or characteristics of the object which is to be detected and/or classified (contrast with background, presence of text or other symbols, closeness of fit to a general template, etc.) as would be understood by those having ordinary skill in the art upon reading the present descriptions.

In some approaches, generating the first representation may include one or more alternative and/or additional suboperations, such as dividing the digital image into a plurality of sections. The digital image may be divided into a plurality of sections in any suitable manner, and in one embodiment the digital image is divided into a plurality of rectangular sections. Of course, sections may be characterized by any shape, and in alternative approaches the plurality of sections may or may not represent the entire digital image, may represent an oversampling of some regions of the image, or may represent a single sampling of each pixel depicted in the digital image. In a preferred embodiment, as discussed above regarding FIGS. 3A-3C, the digital image is divided into S substantially square sections 304 to form an s×s grid 302.

In further approaches, generating the first representation may also include determining, for each section of the digital image, at least one characteristic value, where each characteristic value corresponds to one or more features descriptive of the section. Within the scope of the present disclosures, any feature that may be expressed as a numerical value is suitable for use in generating the first representation, e.g. an average brightness or intensity (0-255) across each pixel in the section, an average value (0-255) of each color channel of each pixel in the section, such as an average red-channel value, and average green-channel value, and an average blue-channel value for a red-green-blue (RGB) image, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

With continuing reference to FIGS. 3A-3C, in some embodiments each pixel 312 of the first representation 310 corresponds to one of the S sections 304 not only with respect to positional correspondence, but also with respect to feature correspondence. For example, in one approach generating the first representation 310 may additionally include determining a characteristic section intensity value i_Sby calculating the average of the individual intensity values i_Pof each pixel 306 in the section 304. Then, each pixel 312 in the first representation 310 is assigned an intensity value equal to the average intensity value i_Scalculated for the corresponding section 304 of the digital image 300. In this manner, the first representation 310 reflects a less granular, normalized representation of the features depicted in digital image 300.

Of course, the pixels 312 comprising the first representation 310 may be represented using any characteristic value or combination of characteristic values without departing from the scope of the presently disclosed classification methods. Further, characteristic values may be computed and/or determined using any suitable means, such as by random selection of a characteristic value from a distribution of values, by a statistical means or measure, such as an average value, a spread of values, a minimum value, a maximum value, a standard deviation of values, a variance of values, or by any other means that would be known to a skilled artisan upon reading the instant descriptions.

In operation 506, a first feature vector is generated based on the first representation.

The first feature vector and/or reference feature matrices may include a plurality of feature vectors, where each feature vector corresponds to a characteristic of a corresponding object class, e.g. a characteristic minimum, maximum, average, etc. brightness in one or more color channels at a particular location (pixel or section), presence of a particular symbol or other reference object at a particular location, dimensions, aspect ratio, pixel density (especially black pixel density, but also pixel density of any other color channel), etc.

As would be understood by one having ordinary skill in the art upon reading the present descriptions, feature vectors suitable for inclusion in first feature vector and/or reference feature matrices comprise any type, number and/or length of feature vectors, such as described in U.S. patent application Ser. No. 12/042,774, filed Mar. 5, 2008; and Ser. No. 12/368,685, filed Feb. 10, 2009 and/or U.S. Pat. No. 7,761,391, granted Jul. 20, 2010 (U.S. patent application Ser. No. 11/752,364, filed May 13, 2007).

In operation 508, the first feature vector is compared to a plurality of reference feature matrices.

The comparing operation 508 may be performed according to any suitable vector matrix comparison, such as described in U.S. patent application Ser. No. 12/042,774, filed Mar. 5, 2008; and Ser. No. 12/368,685, filed Feb. 10, 2009 and U.S. Pat. No. 7,761,391, granted Jul. 20, 2010 (U.S. patent application Ser. No. 11/752,364, filed May 13, 2007).

Thus, in such approaches the comparing may include an N-dimensional feature space comparison. In at least one approach, N is greater than 50, but of course. N may be any value sufficiently large to ensure robust classification of objects into a single, correct object class, which those having ordinary skill in the art reading the present descriptions will appreciate to vary according to many factors, such as the complexity of the object, the similarity or distinctness between object classes, the number of object classes, etc.

As understood herein, “objects” include any tangible thing represented in an image and which may be described according to at least one unique object feature or characteristic such as color, size, dimensions, shape, texture, brightness, intensity, presence or absence of one or more representative mark(s) or other features, location of one or more representative mark(s) or other features, positional relationship between a plurality of representative mark(s) or other features, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. Additionally, objects include any tangible thing represented in an image and which may be classified according to at least one unique combination of such characteristics. For example, in various embodiments objects may include but are in no way limited to persons, animals, vehicles, buildings, landmarks, documents, furniture, plants, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

For example, in one embodiment where attempting to classify an object depicted in a digital image as one of only a small number of object classes (e.g. 3-5 object classes), each object class being characterized by a significant number of starkly distinguishing features or feature vectors (e.g. each object class corresponding to an object or object(s) characterized by very different size, shape, color profile and/or color scheme and easily distinguishable reference symbols positioned in unique locations on each object class, etc.), a relatively low value of N may be sufficiently large to ensure robust classification.

On the other hand, where attempting to classify an object depicted in a digital image as one of a large number of object classes (e.g. 30 or more object classes), and each object class is characterized by a significant number of similar features or feature vectors, and only a few distinguishing features or feature vectors, a relatively high value of N may be preferable to ensure robust classification. Similarly, the value of N is preferably chosen or determined such that the classification is not only robust, but also computationally efficient; i.e. the classification process(es) introduce only minimal processing overhead to the device(s) or system(s) utilized to perform the classification algorithm.

The value of N that achieves the desired balance between classification robustness and processing overhead will depend on many factors such as described above and others that would be appreciated by one having ordinary skill in the art upon reading the present descriptions. Moreover, determining the appropriate value of N to achieve the desired balance may be accomplished using any known method or equivalent thereof as understood by a skilled artisan upon reading the instant disclosures.

In a concrete implementation, directed to classifying driver licenses according to state and distinguishing driver licenses from myriad other document types, it was determined that a 625-dimensional comparison (N=625) provided a preferably robust classification without introducing unsatisfactorily high overhead to processing performed using a variety of current-generation mobile devices.

In operation 510, an object depicted in the digital image is classified as a member of a particular object class based at least in part on the comparing operation 508. More specifically, the comparing operation 508 may involve evaluating each feature vector of each reference list of feature vectors and/or feature matrix, or alternatively evaluating a plurality of feature matrices for objects belonging to a particular object class, and identifying a hyper-plane in the N-dimensional feature space that separates the feature vectors of one reference list of feature vectors and/or feature matrix from the feature vectors of other reference feature matrices. In this manner, the classification algorithm defines concrete hyper-plane boundaries between object classes, and may assign an unknown object to a particular object class based on similarity of feature vectors to the particular object class and/or dissimilarity to other reference list of feature vectors and/or feature matrix profiles.

In the simplest example of such feature-space discrimination, imagining a two-dimensional feature space with one feature plotted along the ordinate axis and another feature plotted along the abscissa, objects belonging to one particular class may be characterized by feature vectors having a distribution of values clustered in the lower-right portion of the feature space, while another class of objects may be characterized by feature vectors exhibiting a distribution of values clustered in the upper-left portion of the feature space, and the classification algorithm may distinguish between the two by identifying a line between each cluster separating the feature space into two classes—“upper left” and “lower-right.” Of course, as the number of dimensions considered in the feature space increases, the complexity of the classification grows rapidly, but also provides significant improvements to classification robustness, as will be appreciated by one having ordinary skill in the art upon reading the present descriptions.

Additional Processing

In some approaches, classification according to embodiments of the presently disclosed methods may include one or more additional and/or alternative features and/or operations, such as described below.

In one embodiment, classification such as described above may additionally and/or alternatively include assigning a confidence value to a plurality of putative object classes based on the comparing operation (e.g. as performed in operation 508 of method 500) the presently disclosed classification methods, systems and/or computer program products may additionally and/or alternatively determine a location of the mobile device, receive location information indicating the location of the mobile device, etc. and based on the determined location, a confidence value of a classification result corresponding to a particular location may be adjusted. For example, if a mobile device is determined to be located in a particular state (e.g. Maryland) based on a GPS signal, then during classification, a confidence value may be adjusted for any object class corresponding to the particular state (e.g. Maryland Driver License, Maryland Department of Motor Vehicle Title/Registration Form, Maryland Traffic Violation Ticket, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions).

Confidence values may be adjusted in any suitable manner, such as increasing a confidence value for any object class corresponding to a particular location, decreasing a confidence value for any object class not corresponding to a particular location, normalizing confidence value(s) based on correspondence/non-correspondence to a particular location, etc. as would be understood by the skilled artisan reading the present disclosures.

The mobile device location may be determined using any known method, and employing hardware components of the mobile device or any other number of devices in communication with the mobile device, such as one or more satellites, wireless communication networks, servers. etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

For example, the mobile device location may be determined based in whole or in part on one or more of a global-positioning system (GPS) signal, a connection to a wireless communication network, a database of known locations (e.g. a contact database, a database associated with a navigational tool such as Google Maps, etc.), a social media tool (e.g. a “check-in” feature such as provided via Facebook, Google Plus, Yelp, etc.), an IP address, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

In more embodiments, classification additionally and/or alternatively includes outputting an indication of the particular object class to a display of the mobile device; and receiving user input via the display of the mobile device in response to outputting the indication. While the user input may be of any known type and relate to any of the herein described features and/or operations, preferably user input relates to confirming, negating or modifying the particular object class to which the object was assigned by the classification algorithm.

The indication may be output to the display in any suitable manner, such as via a push notification, text message, display window on the display of the mobile device, email, etc. as would be understood by one having ordinary skill in the art. Moreover, the user input may take any form and be received in any known manner, such as detecting a user tapping or pressing on a portion of the mobile device display (e.g. by detecting changes in resistance, capacitance on a touch-screen device, by detecting user interaction with one or more buttons or switches of the mobile device, etc.).

In one embodiment, classification further includes determining one or more object features of a classified object based at least in part on the particular object class. Thus, classification may include determining such object features using any suitable mechanism or approach, such as receiving an object class identification code and using the object class identification code as a query and/or to perform a lookup in a database of object features organized according to object class and keyed, hashed, indexed, etc. to the object class identification code.

Object features within the scope of the present disclosures may include any feature capable of being recognized in a digital image, and preferably any feature capable of being expressed in a numerical format (whether scalar, vector, or otherwise), e.g. location of subregion containing reference object(s) (especially in one or more object orientation states, such as landscape, portrait, etc.) object color profile, or color scheme, object subregion color profile or color scheme, location of text. etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

In accordance with another inventive embodiment commensurate in scope with the present disclosures, as shown in FIG. 6, a method 600 is shown. The method 600 may be carried out in any desired environment, and may include embodiments and/or approaches described in relation to FIGS. 1-4D, among others. Of course, more or less operations than those shown in FIG. 6 may be performed in accordance method 600 as would be appreciated by one of ordinary skill in the art upon reading the present descriptions.

In operation 602, a first feature vector is generated based on a digital image captured by a mobile device.

In operation 604, the first feature vector is compared to a plurality of reference feature matrices.

In operation 606, an object depicted in the digital image is classified as a member of a particular object class based at least in part on the comparing (e.g. the comparing performed in operation 604).

In operation 608, one or more object features of the object are determined based at least in part on the particular object class.

In operation 610, a processing operation is performed. The processing operation includes performing one or more of the following subprocesses: detecting the object depicted in the digital image based at least in part on the one or more object features; rectangularizing the object depicted in the digital image based at least in part on the one or more object features; cropping the digital image based at least in part on the one or more object features; and binarizing the digital image based at least in part on the one or more object features.

As will be further appreciated by one having ordinary skill in the art upon reading the above descriptions of document classification, in various embodiments it may be advantageous to perform one or more additional processing operations, such as the subprocesses described above with reference to operation 610, on a digital image based at least in part on object features determined via document classification.

For example, after classifying an object depicted in a digital image, such as a document, it may be possible to refine other processing parameters, functions, etc. and/or utilize information known to be true for the class of objects to which the classified object belongs, such as object shape, size, dimensions, location of regions of interest on and/or in the object, such as regions depicting one or more symbols, patterns, text, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

Regarding performing page detection based on classification, it may be advantageous in some approaches to utilize information known about an object belonging to a particular object class in order to improve object detection capabilities. For example, and as would be appreciated by one having ordinary skill in the art, it may be less computationally expensive, and/or may result in a higher-confidence or higher-quality result to narrow a set of characteristics that may potentially identify an object in a digital image to one or a few discrete, known characteristics, and simply search for those characteristic(s).

Exemplary characteristics that may be utilized to improve object detection may include characteristics such as object dimensions, object shape, object color, one or more reference features of the object class (such as reference symbols positioned in a known location of a document).

In another approach, object detection may be improved based on the one or more known characteristics by facilitating an object detection algorithm distinguishing regions of the digital image depicting the object from regions of the digital image depicting other objects, image background, artifacts, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. For example, if objects belonging to a particular object class are known to exhibit a particular color profile or scheme, it may be simpler and/or more reliable to attempt detecting the particular color profile or scheme within the digital image rather than detecting a transition from one color profile or scheme (e.g. a background color profile or scheme) to another color profile or scheme (e.g. the object color profile or scheme), especially if the two colors profiles or schemes are not characterized by sharply contrasting features.

Regarding performing rectangularization based on classification, it may be advantageous in some approaches to utilize information known about an object belonging to a particular object class in order to improve object rectangularization capabilities. For example, and as would be appreciated by one having ordinary skill in the art, it may be less computationally expensive, and/or may result in a higher-confidence or higher-quality result to transform a digital representation of an object from a native appearance to a true configuration based on a set of known object characteristics that definitively represent the true object configuration, rather than attempting to estimate the true object configuration from the native appearance and project the native appearance onto an estimated object configuration.

In one approach, the classification may identify known dimensions of the object, and based on these known dimensions the digital image may be rectangularized to transform a distorted representation of the object in the digital image into an undistorted representation (e.g. by removing projective effects introduced in the process of capturing the image using a camera of a mobile device rather than a traditional flat-bed scanner, paper-feed scanner or other similar multifunction peripheral (MFP)).

Regarding performing cropping based on classification, and similar to the principles discussed above regarding rectangularization, it may be advantageous in some approaches to utilize information known about an object belonging to a particular object class to improve cropping of digital images depicting the object such that all or significantly all of the cropped image depicts the object and not image background (or other objects, artifacts, etc. depicted in the image).

As a simple example, it may be advantageous to determine an object's known size, dimensions, configuration, etc. according to the object classification and utilize this information to identify a region of the image depicting the object from regions of the image not depicting the object, and define crop lines surrounding the object to remove the regions of the image not depicting the object.

Regarding performing binarization based on classification, the presently disclosed classification algorithms provide several useful improvements to mobile image processing. Several exemplary embodiments of such improvements will now be described with reference to FIGS. 4A-4D.

For example, binarization algorithms generally transform a multi-tonal digital image (e.g. grayscale, color, or any other image such as image 400 exhibiting more than two tones) into a bitonal image, i.e. an image exhibiting only two tones (typically white and black). Those having ordinary skill in the art will appreciate that attempting to binarize a digital image depicting an object with regions exhibiting two or more distinct color profiles and/or color schemes (e.g. a region depicting a color photograph 402 as compared to a region depicting a black/white text region 404, a color-text region 406, a symbol 408 such as a reference object, watermark, etc. object background region 410, etc.) may produce an unsuccessful or unsatisfactory result.

As one explanation, these difficulties may be at least partially due to the differences between the color profiles, schemes, etc., which counter-influence a single binarization transform. Thus, providing an ability to distinguish each of these regions having disparate color schemes or profiles and define separate binarization parameters for each may greatly improve the quality of the resulting bitonal image as a whole and with particular respect to the quality of the transformation in each respective region.

According to one exemplary embodiment shown in FIGS. 4A-4B, improved binarization may include determining an object class color profile and/or scheme (e.g. determining a color profile and/or color scheme for object background region 410); adjusting one or more binarization parameters based on the object class color profile and/or color scheme; and thresholding the digital image using the one or more adjusted binarization parameters.

Binarization parameters may include any parameter of any suitable binarization process as would be appreciated by those having ordinary skill in the art reading the present descriptions, and binarization parameters may be adjusted according to any suitable methodology. For example, with respect to adjusting binarization parameters based on an object class color profile and/or color scheme, binarization parameters may be adjusted to over- and/or under-emphasize a contribution of one or more color channels, intensities, etc. in accordance with the object class color profile/scheme (such as under-emphasizing the red channel for an object class color profile/scheme relatively saturated by red hue(s), etc.).

Similarly, in other embodiments such as particularly shown in FIGS. 4B-4D, improved binarization may include determining an object class mask, applying the object class mask to the digital image and thresholding a subregion of the digital image based on the object class mask. The object class mask may be any type of mask, with the condition that the object class mask provides information regarding the location of particular regions of interest characteristic to objects belonging to the class (such as a region depicting a color photograph 402, a region depicting a black/white text region 404, a color-text region 406, a symbol region depicting a symbol 408 such as a reference object, watermark, etc., an object background region 410, etc.) and enabling the selective inclusion and/or exclusion of such regions from the binarization operation(s).

For example, as shown in FIG. 4B, improved binarization includes determining an object class mask 420 identifying regions such as discussed immediately above and applying the object class mask 420 to exclude from binarization all of the digital image 400 except a single region of interest, such as object background region 410. Alternatively the entire digital image may be masked-out and a region of interest such as object background region 410 subsequently masked-in to the binarization process. Moreover, in either event the masking functionality now described with reference to FIG. 4B may be combined with the exemplary color profile and/or color scheme information functionality described above, for example by obtaining both the object class mask and the object color profile and/or color scheme, applying the object class mask to exclude all of the digital image from binarization except object background region 410, adjusting one or more binarization parameters based on the object background region color profile and/or color scheme, and thresholding the object background region 410 using the adjusted binarization parameters.

Extending the principle shown in FIG. 4B, multiple regions of interest may be masked-in and/or masked-out using object class mask 420 to selectively designate regions and/or parameters for binarization in a layered approach designed to produce high-quality bitonal images. For example, as shown in FIG. 4C multiple text regions 404, 406 may be retained for binarization (potentially using adjusted parameters) after applying object class mask 420, for example to exclude all non-text regions from binarization, in some approaches.

Similarly, it may be advantageous to simply exclude only a portion of an image from binarization, whether or not adjusting any parameters. For example, with reference to FIG. 4D, it may be desirable to mask-out a unique region of a digital image 400, such as a region depicting a color photograph 402, using an object class mask 420. Then, particularly in approaches where the remaining portion of the digital image 400 is characterized by a single color profile and/or color scheme, or a small number (i.e. no more than 3) substantially similar color profile and/or color schemes, binarization may be performed to clarify the remaining portions of the digital image 400. Subsequently, the masked-out unique region may optionally be restored to the digital image 400, with the result being an improved bitonal image quality in all regions of the digital image 400 that were subjected to binarization coupled with an undisturbed color photograph 402 in the region of the image not subjected to binarization.

In still more embodiments, it may be advantageous to perform optical character recognition (OCR) based at least in part on the classification and/or result of classification. Specifically, it may be advantageous to determine information about the location, format, and/or content of text depicted in objects belonging to a particular class, and modify predictions estimated by traditional OCR methods based on an expected text location, format and/or content. For example, in one embodiment where an OCR prediction estimates text in a region corresponding to a “date” field of a document reads “Jan. 14, 2011” the presently disclosed algorithms may determine the expected format for this text follows a format such as “[Abbreviated Month][.][##][,][####]” the algorithm may correct the erroneous OCR predictions, e.g. converting the comma after “Jan” into a period and/or converting the letter “l” at the end of 2011” into a numerical one character. Similarly, the presently disclosed algorithms may determine the expected format for the same text is instead “[##]/[##]/[####]” and convert “Jan” to “01” and convert each set of comma-space characters “,” into a slash “/” to correct the erroneous OCR predictions.

Of course, other methods of improving upon and/or correcting OCR predictions that would be appreciated by the skilled artisan upon reading these descriptions are also fully within the scope of the present disclosure.

Data Extraction

In addition to performing improved image processing based on object classification, a user may wish to gather information about one or more objects depicted in a digital image. In several embodiments, it is advantageous to leverage object classification for the purposes of extracting data from digital images. As described in further detail below, the presently disclosed methods, systems and computer program products thus include functionality for extracting data from digital images based on object classification.

The data extraction embodiments discussed herein may utilize one or more of the functionalities disclosed in related U.S. patent application Ser. No. 12/042,774, filed Mar. 5, 2008; and Ser. No. 12/368,685, filed Feb. 10, 2009, and U.S. Pat. No. 7,761,391, granted Jul. 20, 2010 (U.S. patent application Ser. No. 11/752,364, filed May 13, 2007) each of which is herein incorporated by reference in its entirety. For example, the presently disclosed embodiments of data extraction may utilize one or more of support-vector-machine (SVM) techniques, learn-by-example (LBE) techniques, feature vectors, feature matrices, document validation techniques, dataset organization techniques, transductive classification techniques, maximum entropy discrimination (MED) techniques, etc. as disclosed therein.

Referring now to FIG. 7, a method 700 is shown, according to one embodiment. The method may be performed in any suitable environment and/or utilizing any suitable mechanism(s), including those depicted in FIGS. 1-4D, in various approaches.

In one approach, method 700 includes operation 702, where a digital image captured by a mobile device is received. The digital image may be received and/or stored in a memory of the mobile device or another mobile device, in some embodiments. Moreover, the digital image may be received from a variety of sources, such as a component of the mobile device including a camera, a memory, wireless receiver, antenna, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. In other approaches, the digital image may be received from a remote device, such as a remote server, another mobile device, a camera with integrated data transmission capability, a facsimile machine or other multifunction printer, etc. The digital image may optionally be received via an online service, a database, etc. as would be understood by the skilled artisan reading this disclosure.

Method 700 further includes performing operations 704-710 using a processor of the mobile device, which are described in further detail below. As will be understood by one having ordinary skill in the art upon reading the present descriptions, various embodiments of method 700 may involve performing any of operations 704-710 using a processor of a mobile device, a processor of a server, a cloud computing environment, etc. as well as any combination thereof.

In operation 704, the mobile device processor is used to determine whether an object depicted in the digital image belongs to a particular object class among a plurality of object classes. Determining whether the depicted object belongs to a particular object class may be performed using any method as described herein, with particular reference to the object classification methods discussed above with reference to FIGS. 5 and 6.

In operation 706, the mobile device processor is used to determine one or more object features of the object based at least in part on the particular object class. In one embodiment, object features may be determined using a feature vector, feature vector list, feature matrix, and/or extraction model. This determination is made in response to determining the object belongs to the particular object class. As discussed herein, object features include any unique characteristic or unique combination of characteristics sufficient to identify an object among a plurality of possible objects or any unique characteristic or unique combination of characteristics sufficient to identify an object as belonging to a particular object class among a plurality of object classes. For example, in various approaches object features may include object color, size, dimensions, shape, texture, brightness, intensity, presence or absence of one or more representative mark(s) or other features, location of one or more representative mark(s) or other features, positional relationship between a plurality of representative mark(s) or other features, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

In a preferred embodiment, one or more object features comprise one or more regions of interest of the object. As understood herein, a region of interest may include any portion of the object that depicts, represents, contains, etc. information the user desires to extract. Thus, in some approaches one or more of the regions of interest comprise one or more text characters, symbols, photographs, images, etc.

For example, in one instance a user may wish to perform a credit check, apply for a loan or lease, etc. In order to perform the desired action, the user needs to gather data, such as an applicant's name, address, social security number, date of birth, etc. The mobile device may receive a digital image of one or more identifying documents such as a utility bill, driver license, social security card, passport, pay stub, etc. which contains/depicts information relevant to performing the credit check, loan or lease application, etc. In this case, regions of interest may include any portion of the identifying document that depicts relevant data, such as the applicant's name, address, social security number, date of birth, etc.

In another example, a user wishes to make an electronic funds transfer, set up a recurring bill payment, engage in a financial transaction, etc. In this case the user may need to gather data such as an account number, routing number, payee name, address, biller name and/or address, signature, payment amount, payment date and/or schedule, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. The mobile device may receive a digital image of one or more financial documents such as a bill, remittance coupon, check, credit card, driver license, social security card, passport, pay stub, etc. which contains/depicts information relevant to performing the credit check, loan or lease application, etc. In this case, regions of interest may include any portion of the identifying document that depicts relevant data, such as the account number, routing number, payee name, address, biller name and/or address, signature, payment amount, payment date and/or schedule, etc.

In still another example, a user wishes to authenticate the identity of an individual applying for motor vehicle registration or a new bank account, etc. The applicant provides a driver license as proof of identification. The user may capture an image of the driver license, and extract data from the image including text information such as a name, address, driver license number, etc. The user may also extract a photograph of the licensee from the image, and compare the extracted photograph to a reference photograph of the licensee. The reference photograph may be retrieved from a local database maintained by the motor vehicle administration office, bank, a database maintained by a government agency, etc. in various approaches. Alternatively, the reference photograph may be a previously obtained photograph of the licensee, for example a photograph obtained during a prior transaction requiring identity authentication. Based on the comparison, the user may be presented an indication of whether the extracted photograph matches the reference photograph, along with an optional confidence score, in one embodiment.

Operation 708 includes using the mobile device processor to build or select an extraction model based at least in part on the one or more object features. In one embodiment, the object class determines the extraction model. As understood herein, an extraction model encompasses any model that may be applied to a digital image in order to extract data therefrom. In a preferred approach, the extraction model comprises a set of instructions and/or parameters for gathering data from a digital image. In a particularly preferred embodiment, the extraction model utilizes a feature vector and/or list of feature vectors and/or feature matrix to generate and/or modify instructions for extracting data from digital images.

For example, in one approach an exemplary data extraction process as described herein is configured to extract data from various forms of identification based on objects and/or object features (as may be embodied in one or more feature vector(s)) thereof. Illustrative forms of identification may include, for example, a plurality of driver's license formats. Moreover, the illustrative IDs may be classified according to one or more distinguishing criteria, such as an issuing entity (state, administrative agency, etc.) to which the ID corresponds. The extraction model may be selected based on determining an ID in question belongs to one of the predetermined categories of ID (e.g. the ID in question is a Maryland driver's license). Preferably, the selected extraction model is built using a plurality of exemplars from the corresponding category/class.

In one exemplary approach, based on the user input identifying the region(s) of interest, operation 708 may include reviewing one or more existing object class definitions to determine whether the determined object features define a pattern that matches, corresponds, or is similar to a pattern defining features of objects belonging to the existing object class. Upon determining the patterns match, correspond or are similar, operation 708 may include selecting an existing extraction model defined for the matching object class, and utilize that extraction model to extract data from the digital image. The existing object class definition and/or extraction model may be retrieved from a memory of the mobile device, a memory in communication with the mobile device, a server, a local or online database, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

Alternatively, operation 708 may include analyzing the image and characteristics thereof to define a feature vector descriptive of the image characteristics. This new feature vector may be used to modify a feature vector, list of feature vectors and/or feature matrix descriptive of the existing object class having the matching, corresponding, or similar pattern of regions of interest. For example, building the extraction model may include mapping the object features to the feature vector, list of feature vectors and/or feature matrix, which may have been modified via the new feature vector as described above. The resulting extraction model is configured to extract data from images depicting objects belonging to the existing object class.

Additionally and/or alternatively, operation 708 may include building a new extraction model based on the object features, in some approaches. More specifically, using the processor of the mobile device, the image is analyzed and characteristics thereof used to define a feature vector and/or list of feature vectors descriptive of the image characteristics. For example, the feature vector(s) may correspond to image characteristics such as pixel brightness and/or intensity in one or more color channels, brightness and/or intensity of one or more neighboring pixels in one or more color channels, positional relationship of pixels in the image or in a subregion of the image, etc. Image analysis and feature vector definition may be performed in any suitable manner, and preferably may be performed substantially as described above regarding “Document Classification” and “Additional Processing.” Using the feature vector(s), operation 708 may include building an extraction model configured to extract data corresponding to image characteristics depicted in the region(s) of interest.

In still further embodiments, building an extraction model may include mapping the feature vector, list of feature vectors, and/or feature matrix and associating metadata label(s) with each mapped object feature. In one approach mapping the feature vector, list of feature vectors, and/or feature matrix to object features involves processing the feature vector to determine therefrom pertinent location information, color profile information, etc. for the image.

Metadata labels may include any type of information and be associated with any type of object feature. For example, in some embodiments metadata labels may identify object features according to type of data depicted, such as text, alphanumeric characters, symbols, numeric characters, picture, background, foreground, field, shadow, texture, shape, dimension, color profile or scheme, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

In case of invoices for instance, metadata labels may include text and/or relative or absolute location information. For example, a metadata label may identify text as an invoice number with an absolute location at a bottom right corner of the invoice. Moreover, another metadata label may identify text as an invoice date with a relative location directly below the invoice number address on the invoice, etc.

Additionally and/or alternatively, metadata labels may identify object features according to relevance in subsequent processing operations, such as identifying particular data format or informational content. For example, metadata labels may include personal information labels such as “name,” “address,” “social security number,” “driver license number,” “date of birth,” “credit score,” “account number,” “routing number,” “photograph” etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

In operation 710, the mobile device processor is used to extract data from the digital image based at least in part on the extraction model. Notably, extracting the data does not utilize optical character recognition (OCR) techniques. However, as described in further detail below, optical character recognition techniques may be utilized outside the context of data extraction as performed in operation 710.

Referring now to FIG. 8, a method 800 is shown, according to one embodiment. The method may be performed in any suitable environment and/or utilizing any suitable mechanism(s), including those depicted in FIGS. 1-4D, in various approaches. In one view, the method 800 may be considered an implementation of a data extraction process as described herein; the implementation is in the format of a mobile application being engaged by a user.

In one approach, method 800 includes operation 802, where a digital image captured by a mobile device is received. The digital image may be received and/or stored in a memory of the mobile device or another mobile device, in some embodiments. Moreover, the digital image may be received from a variety of sources, such as a component of the mobile device including a camera, a memory, wireless receiver, antenna, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. In other approaches, the digital image may be received from a remote device, such as a remote server, another mobile device, a camera with integrated data transmission capability, a facsimile machine or other multifunction printer, etc. The digital image may optionally be received via an online service, a database, etc. as would be understood by the skilled artisan reading this disclosure.

Method 800 further includes performing operations 804-812 using one or more of a processor of the mobile device, a processor of a server, one or more processors of a remote cloud computing environment, etc., which are described in further detail below.

In operation 804, the mobile device processor is used to determine whether an object depicted in the digital image belongs to a particular object class among a plurality of object classes. Determining whether an object belongs to a particular object class may be performed according to any suitable method, and is preferably performed in a manner commensurate with the descriptions above regarding Document Classification, e.g. as set forth with respect to FIGS. 5 & 6, in various embodiments.

In operation 806, once again using the processor of the mobile device, the digital image is displayed on a mobile device display. The digital image is displayed in response to determining the object does not belong to any particular object class among the plurality of object classes. Additionally and/or alternatively, the digital image may be displayed in response to determining the object does belong to a particular object class among the plurality of object classes.

Displaying the digital image on the mobile device display enables further action conducive to efficient and robust extraction of data from digital images using a processor. For example, in various approaches the digital image may be displayed on the mobile device display to provide feedback regarding the digital image, such as image quality, object classification (or lack thereof), extracted data, etc. Similarly, the digital image may be displayed to facilitate receiving additional input from a user, such as: user feedback regarding a classification and/or extraction result; metadata associated with or to be associated with the digital image, object depicted therein, and/or a particular object class to which the depicted object is determined to belong, etc.; instructions to perform additional processing, extraction, or other manipulation of the digital image, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

Operation 808 includes using the mobile device processor to receive user input via the display of the mobile device. More particularly, the user input identifies one or more regions of interest in the object. In one embodiment of a method 800 including operation 808, the image of the identifying document may be presented to the user via the mobile device display. The user may be prompted to confirm, negate, and/or modify regions of interest identified based on an object classification. The user may additionally and/or alternatively be prompted to define, confirm, negate, and/or modify one or more regions of interest not identified based on the classification, in various embodiments.

In operation 810, an extraction model is built and/or selected based at least in part on the user input received in operation 808. In one exemplary approach, based on the user input identifying the region(s) of interest, operation 810 may include reviewing one or more existing object class definitions to determine whether the identified regions of interest define a pattern that matches, corresponds, or is similar to a pattern defining regions of interest of objects belonging to the existing object class. Upon determining the patterns match, correspond or are similar, operation 810 may include selecting an existing extraction model defined for the matching object class, and utilize that extraction model to extract data from the digital image. Alternatively, operation 810 may include analyzing the image and characteristics thereof to define a feature vector descriptive of the image characteristics. This new feature vector may be used to modify a list of feature vectors and/or feature matrix descriptive of the existing object class having the matching, corresponding, or similar pattern of regions of interest. The resulting extraction model is configured to extract data from images depicting objects belonging to the existing object class, including raw image data, data corresponding to text, images, photographs, symbols, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions.

Additionally and/or alternatively, operation 810 may include building a new extraction model based on the user input defining the regions of interest, in some approaches. More specifically, using the processor of the mobile device, the image is analyzed and characteristics thereof used to define a feature vector descriptive of the image characteristics. For example, the feature vector may correspond to image characteristics such as pixel brightness and/or intensity in one or more color channels, brightness and/or intensity of one or more neighboring pixels in one or more color channels, positional relationship of pixels in the image or in a subregion of the image, regions of a document likely to depict text, regions of a document likely to depict a photograph, etc. Image analysis and feature vector definition may be performed in any suitable manner, and preferably may be performed substantially as described above regarding “Document Classification” and “Additional Processing.” Using the feature vector, operation 810 may include building an extraction model configured to extract data corresponding to image characteristics depicted in the region(s) of interest.

In operation 812, data is extracted from the image based at least in part on the extraction model. Notably, the extraction process does not utilize any OCR techniques. Rather, the extraction model is defined based on a feature vector list of feature vectors and/or feature matrix descriptive of an object or object class, respectively.

In one illustrative embodiment, a user, via a mobile application adapted to facilitate performing data classification and/or extraction as described herein, may perform a classification operation to attempt classifying an object depicted in a digital image. Depending on whether the classification algorithm has been trained to recognize an object as belonging to a particular object class, the algorithm may or may not successfully classify the particular object depicted in the digital image. After completing one or more classification attempts, the image of the identifying document may be presented to the user via the mobile device display. The user may be prompted to confirm, negate, and/or modify regions of interest identified based on the object classification. The user may additionally and/or alternatively be prompted to define one or more regions of interest not identified based on the classification.

Similarly, if the classification attempt fails to identify the object class, the user may be prompted to define a new object class and further define one or more regions of interest in object belonging to the new object class by interacting with the display of the mobile device. For example, a user may draw one or more bounding boxes around regions of interest by providing tactile feedback via the mobile device display. The user may then direct the mobile application to extract data from the digital image, and the application optionally employs the processor of the mobile device, a server, etc. to build and/or select an extraction model based at least in part on the user-defined regions of interest and extract data from the digital image based in whole or in part on the extraction model.

In various approaches, methods 700 and/or 800 may optionally include one or more additional functionalities, features and/or operations as described below.

In one approach, method 700 and/or method 800 may further include training the extraction model. Training an extraction model may be accomplished using any known method, model, mechanism, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. In a preferred embodiment, training comprises a learn-by-example (LBE) process. Specifically, for a particular object class, a plurality of representative objects may be provided with or without associated metadata labels. Based at least in part on the object features of the provided representative objects, the extraction model may be trained to modify and thus improve the robustness of extracting data from objects belonging to the object class.

Those having ordinary skill in the art will appreciate that in some approaches extraction model training may be specifically designed to improve the ability of the extraction model to precisely and accurately extract data from objects corresponding to a particular object class for which the extraction model was built. Such training may improve extraction precision and/or accuracy by training the model by providing a set of objects characterized by substantially identical object features, e.g. a plurality of copies of the same object type such as a standardized form, document type, multiple images of the same object, etc. Using this training set, the extraction model may reinforce the list of feature vectors and/or feature matrix representing objects in the class, and improve the robustness of extracting data from objects belonging to the class.

Alternatively, training may be specifically designed to improve the ability of the extraction model to extract data from a set of objects within an object class characterized by variable object features, or across several object classes. Such training may involve providing a set of objects with diverse characteristics to improve the extraction model's ability to extract data generally from a diverse object class or several object classes.

In more approaches, the extraction model may be trained using the processor of the mobile device. Moreover, the trained extraction model may be stored and/or exported, e.g. to a memory, a buffer, another process or processor, etc. The trained extraction model may be preferably stored and/or exported to a memory of the mobile device, a processor of the mobile device, or another process being executed using the processor of the mobile device. In various embodiments, the trained extraction model may be flagged and/or retrieved for subsequent use by the mobile device or another mobile device. Similarly, the extraction model may be stored and/or passed to a memory and/or processor of another device, such as another mobile device, a server, a cloud computing environment, etc.

Preferably, training as described herein utilizes a training set comprising a plurality of objects, and more preferably the training set comprises no less than five objects.

In addition to training the extraction model, some embodiments of method 700 may additionally and/or alternatively encompass performing at least one OCR technique on one or more regions of the digital image. The OCR'ed region(s) may correspond to one or more of the object features (e.g. object features identified using the object class definition and/or extraction model) and/or other object features (e.g., features were not previously identified using the object class definition and/or extraction model).

Still more embodiments of method 700 and/or method 800 may further include detecting one or more lines of text in objects such as documents. In some approaches, detecting text lines involves projecting the digital image onto a single dimension. In an exemplary approach, a projection may be made along a dimension perpendicular to the predominant axis of text line orientation, so that text lines and areas between lines of text can be easily distinguished according to dark area density (e.g. black pixel density, count, etc.). Thus, if the document is oriented in portrait, detecting text includes projecting along the vertical dimension (y-axis) and if the document is in landscape orientation, detecting text includes projecting along the horizontal dimension (x-axis). The projection can also be used to determine and/or manipulate orientation (portrait, landscape, or any other angle of skew) of a document, in other approaches. For example, in an exemplary embodiment configured to classify and extract data from images of documents corresponding to a standard ID such as a driver's license, detecting lines of text, etc. may be utilized to determine a most likely document orientation from among a plurality of possible orientations (e.g. 0°, 90°, 180°, or 270° rotation angle, in one approach).

Detecting text lines may additionally include determining a distribution of light and dark areas along the projection; determining a plurality of dark pixel densities. Moreover, each dark pixel density may correspond to a position along the projection. Upon determining the plurality of dark pixel densities, probable text lines may be determined according to whether the corresponding dark pixel density is greater than a probable text line threshold, which may be pre-defined by a user, determined experimentally, automatically determined, etc. In embodiments where a probable text line threshold is employed, detecting text lines further includes designating each position as a text line upon determining the corresponding dark pixel density is greater than the probable text line threshold.

In another embodiment detecting text lines may include detecting a plurality of connected components non-background elements in the digital image, and determining a plurality of likely characters based on the plurality of connected components. Likely characters may be regions of a digital image characterized by a predetermined number of light-to-dark transitions in a given direction, such as three light-to-dark transitions in a vertical direction as would be encountered for a small region of the digital image depicting a capital letter “E,” each light-to-dark transition corresponding to a transition from a background of a document (light) to one of the horizontal strokes of the letter “E.” Of course, other numbers of light-to-dark transitions may be employed, such as two vertical and/or horizontal light-to-dark transitions for a letter “o,” one vertical light to dark transition for a letter “l,” etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions. Upon determining likely characters, lines of text may be determined by identifying regions of the image having a plurality of adjacent characters, and may define the text lines according to a text baseline of the plurality of adjacent characters, in one embodiment.

In some approaches, it is possible to determine and/or manipulate image orientation based on the result of projecting the image along one dimension. For example, if a projection produces a one-dimensional pattern of high dark pixel density regions interspersed with low dark pixel density regions, then it is probable that the projection was made along the axis perpendicular to the longitudinal axis of text line orientation (i.e. a projection along the y-axis for a document in “portrait” orientation having text oriented from left-to-right along the x-axis of the image; or a projection along the x-axis for a document in “landscape” orientation having text oriented bottom-to-top along the y-axis of the image). Based on this probabilistic determination, one may optionally rotate and/or reorient the image based on projection results.

Those having ordinary skill in the art will appreciate that detecting text lines based on dark pixel density along a one-dimensional projection may be particularly challenging for color images. For example, some documents may depict text in a color relatively lighter than the background, in which case dark pixel density would indicate absence of a probable text line rather than presence thereof. In that case, detecting text lines may involve designating probable text lines near any position along the projection upon determining the position is characterized by a dark pixel density less than a dark pixel density threshold.

Moreover, text may be presented in a variety of colors, and mere dark pixel density may be an insufficient characteristic from which to identify probable text lines. In order to address these challenges, detecting text lines in color images preferably includes projecting each color channel of the digital image onto a single channel along the single dimension. In other words, color channel intensity values (e.g. integer values between 0-255) are transformed into a single intensity value. The transformation may be accomplished according to any suitable function, and in a preferred embodiment the intensity of each color channel for a given pixel or set of pixels is averaged and the pixel or set of pixels is assigned a representative intensity value according to the average of the color channel intensity values.

In another embodiment, data extraction may include associating object classes with one or more lists of object regions containing information of interest, for example a list of rectangular regions of a document that contain text, or that may contain text and the color of expected text.

In one illustrative example, a user is presented an image of an object via a display of a mobile device. The user interacts with the image via the mobile device display, and defines one or more regions of interest, for example indicating a region displaying the user's name, address, license number, etc. The user can repeat the process for a multitude of images, and thus provide training images either intentionally or as part of a transparent process. Once a sufficient number of training examples have been defined (e.g. about 5 in the case of a small document such as a driver license), the training algorithm may be executed automatically or at the user's discretion. The result of training is an extraction model can be used to automatically extract the relevant locations and rectangles of subsequently presented unknown documents, all without utilizing OCR techniques.

In other approaches, after data is extracted according to the above-described methods, OCR techniques may be utilized for purposes other than mere data extraction. For example, OCR may be performed using a processor of the mobile device, and may only OCR a small subset of the total image. Alternatively, OCR may be performed using a processor of a server. In order reduce communication time between the mobile device and the server, only portion(s) of the image to be processed using OCR may be transmitted to the server.

In additional embodiments, classification and/or extraction results may be presented to a user for validation, e.g. for confirmation, negation, modification of the assigned class, etc. For example, upon classifying an object using semi- or fully automated processes in conjunction with distinguishing criteria such as defined herein, the classification and the digital image to which the classification relates may be displayed to a user (e.g. on a mobile device display) so that the user may confirm or negate the classification. Upon negating the classification, a user may manually define the “proper” classification of the object depicted in the digital image. This user input may be utilized to provide ongoing “training” to the classifier(s), in preferred approaches. Of course, user input may be provided in relation to any number of operations described herein without departing from the scope of the instant disclosures.

In even more preferred embodiments, the aforementioned validation may be performed without requiring user input. For instance, it is possible to mitigate the need for a user to review and/or to correct extraction results by performing automatic validation of extraction results. In general, this technique involves referencing an external system or database in order to confirm whether the extracted values are known to be correct. For example, if name and address are extracted, in some instances it is possible to validate that the individual in question in fact resides at the given address.

This validation principle extends to classification, in even more embodiments. For example, if the extraction is correct, in some approaches it is appropriate to infer that the classification is also correct. This inference relies on the assumption that the only manner in which to achieve the “correct” extraction result (e.g. a value matches an expected value in a reference data source, matches an expected format for the value in question, is associated with an expected symbol or other value, etc. as would be understood by one having ordinary skill in the art upon reading the present descriptions).

While the present descriptions of data extraction within the scope of the instant disclosure have been made with primary reference to methods, one having ordinary skill in the art will appreciate that the inventive concepts described herein may be equally implemented in or as a system and/or computer program product.

For example, a system within the scope of the present descriptions may include a processor and logic in and/or executable by the processor to cause the processor to perform steps of a method as described herein, such as methods 700 and 800.

Similarly, a computer program product within the scope of the present descriptions may include a computer readable storage medium having program code embodied therewith, the program code readable/executable by a processor to cause the processor to perform steps of a method as described herein, such as methods 700 and 800.

The inventive concepts disclosed herein have been presented by way of example to illustrate the myriad features thereof in a plurality of illustrative scenarios, embodiments, and/or implementations. It should be appreciated that the concepts generally disclosed are to be considered as modular, and may be implemented in any combination, permutation, or synthesis thereof. In addition, any modification, alteration, or equivalent of the presently disclosed features, functions, and concepts that would be appreciated by a person having ordinary skill in the art upon reading the instant descriptions should also be considered within the scope of this disclosure.

Accordingly, one embodiment of the present invention includes all of the features disclosed herein, including those shown and described in conjunction with any of the FIGS. Other embodiments include subsets of the features disclosed herein and/or shown and described in conjunction with any of the FIGS. Such features, or subsets thereof, may be combined in any way using known techniques that would become apparent to one skilled in the art after reading the present description.

While various embodiments have been described above, it should be understood that they have been presented by way of example only, and not limitation. Thus, the breadth and scope of an embodiment of the present invention should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

Claims

1. A method, comprising: receiving or capturing digital image data;detecting an object depicted in a digital image represented by the digital image data, wherein detecting the object comprises defining a plurality of candidate edge points based on statistically significant differences between each respective candidate edge point and a background of the digital image;processing the digital image to improve an image quality of at least the object;determining an object classification based on the processed digital image;building or determining an extraction model based at least in part on the object classification; andextracting, using the extraction model, information from one or more regions of interest within the digital image;wherein each region of interest depicts one or more object features selected from: text characters, symbols, photographs, and image characteristics; andwherein the information comprises one or more of: an object feature color;an object feature size;an object feature dimensions;an object feature shape;an object feature texture;an object feature brightness;an object feature intensity in one or more color channels;a position of one or more object features depicted in the digital image; anda positional relationship between one or more object features depicted in the digital image.
2. The method as recited in claim 1, wherein the detecting is based on the one or more object features.
3. The method as recited in claim 1, further comprising associating metadata labels to one or more of the object features, wherein the metadata labels identify one or more of: an object feature data type;an object feature data format;an object feature text information content; andan object feature location information.
4. The method as recited in claim 1, wherein the processing comprises one or more of: rectangularizing the object based at least in part on the one or more object features;cropping the digital image based at least in part on the one or more object features; andbinarizing the digital image based at least in part on the one or more object features.
5. The method as recited in claim 1, wherein the detecting comprises a window-based analysis of a plurality of sections of the digital image.
6. The method as recited in claim 5, wherein defining each candidate edge point comprises: defining one or more large analysis windows within the digital image;defining a plurality of small analysis windows within the digital image;estimating one or more distributions of statistics for each large analysis window;calculating one or more statistics for each small analysis window;determining whether a statistically significant difference exists between one or more of the statistics calculated for each of the small analysis windows and a corresponding distribution of statistics estimated for the one or more large analysis windows; anddesignating a point in each small analysis window for which the statistically significant difference exists as a candidate edge point upon determining the statistically significant difference exists.
7. The method as recited in claim 5, wherein the window-based analysis employs a plurality of large analysis windows several pluralities of small analysis windows, each plurality of small analysis windows corresponding to one of the large analysis windows.
8. The method as recited in claim 5, wherein the window-based analysis comprises a step-over approach analyzing the digital image according to a predefined pattern.
9. The method as recited in claim 8, wherein the predefined pattern is selected from: spiral; serpentine; shingled; checkerboard; and striped.
10. The method as recited in claim 1, wherein the statistically significant differences are evaluated according to one or more metrics selected from: a p-value, a z-test, and a chi-squared correlation.
11. The method as recited in claim 1, wherein the processing comprises correcting one or more of distortion and skew within the digital image, and wherein the correcting comprises: generating a rectangularized digital image from the digital image;determining a pair of rectangle-based coordinates for each point within the rectangularized digital image; andmatching each pair of rectangle-based coordinates to an equivalent pair of tetragon-based intrinsic coordinates within the digital image.
12. The method as recited in claim 1, wherein the processing comprises: detecting uneven illumination within the digital image, andnormalizing illumination within the digital image in response to detecting the uneven illumination.
13. The method as recited in claim 12, wherein detecting the uneven illumination comprises: dividing the digital image into a plurality of sections;determining whether each section is oversaturated;determining whether each section is undersaturated;determining a number of oversaturated sections;determining a number of undersaturated sections;determining that the digital image is oversaturated upon determining that a ratio of the number of oversaturated sections to a total number of sections exceeds an oversaturation threshold;determining that the digital image is undersaturated upon determining that a ratio of the number of undersaturated sections to the total number of sections exceeds an undersaturation threshold; anddetermining that an illumination problem exists in the digital image upon determining that the digital image is either undersaturated or oversaturated; andwherein normalizing the illumination comprises: deriving a two-dimensional illumination model from the digital image; andapplying the two-dimensional illumination model to each pixel in the digital image.
14. The method as recited in claim 1, comprising determining one or more object features of the object based on a feature vector corresponding to the object classification.
15. The method as recited in claim 1, wherein the processing comprises detecting blur within the digital image; and wherein detecting the blur comprises: dividing the digital image into a plurality of sections, each section comprising a plurality of pixels, and, for each section: determining whether the section contains one or more sharp pixel-to-pixel transitions in a first direction;counting a total number of first direction sharp pixel-to-pixel transitions for the section (SS1);determining whether the section contains one or more blurred pixel-to-pixel transitions in the first direction;counting a total number of first-direction blurred pixel-to-pixel transitions for the section (SB1);determining whether the section contains one or more sharp pixel-to-pixel transitions in a second direction;counting a total number of second-direction sharp pixel-to-pixel transitions for the section (SS2);determining whether the section contains one or more blurred pixel-to-pixel transitions in the second direction;counting a total number of second-direction blurred pixel-to-pixel transitions for the section (SB2);determining the section is blank upon determining: SS1 is less than a predetermined sharp transition threshold,SB1 is less than a predetermined blurred transition threshold,SS2 is less than a predetermined sharp transition threshold, andSB2 is less than a predetermined blurred transition threshold; anddetermining, for all non-blank sections, a first direction blur ratio r1=SS1/SB1;determining, for all non-blank sections, a second direction blur ratio r2=SS2/SB2;determining that a non-blank section is blurred in the first direction upon determining that r1 is less than a predefined section blur ratio threshold; anddetermining that a non-blank section is blurred in the second direction upon determining that r2 is less than the predefined section blur ratio threshold; anddetermining that a non-blank section is bluffed upon determining one or more of: the section is blurred in the first direction, andthe section is blurred in the section direction; anddetermining a total number of blurred sections;calculating an image blur ratio R comprising the total number blurred sections to a total number of sections; anddetermining the digital image is blurred upon determining the image blur ratio is greater than a predetermined image blur threshold.
16. The method as recited in claim 1, wherein the processing comprises: estimating a resolution of the digital image; and wherein estimating the resolution of the digital image comprises: detecting a plurality of connected components of a plurality of non-background elements in the digital image;determining a plurality of likely characters based on the plurality of connected components;determining one or more average character dimensions based on the plurality of likely characters; andestimating the resolution of the digital image based on the one or more average character dimensions.
17. The method as recited in claim 1, wherein the information further comprises either or both of: an object feature text information; and an object feature text format.
18. A mobile device configured to execute logic, the mobile device comprising logic configured to cause the mobile device, upon execution thereof, to: receive or capture digital image data;detect a document depicted in a digital image represented by the digital image data, wherein detecting the document comprises defining a plurality of candidate edge points based on statistically significant differences between each respective candidate edge point and a background of the digital image;process the digital image to improve an image quality of at least the document;determine a document classification based on the processed digital image;build or determine an extraction model based at least in part on the document classification; andextract, using the extraction model, information from one or more regions of interest within the document;wherein each region of interest depicts one or more object features selected from: text characters, symbols, photographs, and image characteristics; andwherein the information comprises one or more of: an object feature color;an object feature size;an object feature dimensions;an object feature shape;an object feature texture;an object feature brightness;an object feature intensity in one or more color channels;a position of one or more object features depicted in the digital image; anda positional relationship between one or more object features depicted in the digital image.
19. A computer program product comprising a non-transitory computer readable storage medium having embodied thereon executable instructions configured to cause a mobile device, upon execution of the instructions, to: receive or capture digital image data;detect a document depicted in a digital image represented by the digital image data, wherein detecting the document comprises defining a plurality of candidate edge points based on statistically significant differences between each respective candidate edge point and a background of the digital image;process the digital image to improve an image quality of at least the document;determine a document classification based on the processed digital image;build or determine an extraction model based at least in part on the document classification; andextract, using the extraction model, information from one or more regions of interest within the document;wherein each region of interest depicts one or more object features selected from: text characters, symbols, photographs, and image characteristics; andwherein the information comprises one or more of: an object feature color;an object feature size;an object feature dimensions;an object feature shape;an object feature texture;an object feature brightness;an object feature intensity in one or more color channels;a position of one or more object features depicted in the digital image; anda positional relationship between one or more object features depicted in the digital image.

RELATED APPLICATIONS

This application is a continuation of U.S. patent application Ser. No. 14/209,825, filed Mar. 13, 2014, which claims priority to U.S. Provisional Patent Application No. 61/780,747, filed Mar. 13, 2013, which are herein incorporated by reference. This application is related to U.S. Pat. No. 8,855,375, which is herein incorporated by reference in its entirety.

US Referenced Citations (562)

Number	Name	Date	Kind
1660102	Appelt et al.	Feb 1928	A
3069654	Hough	Dec 1962	A
3696599	Palmer et al.	Oct 1972	A
4558461	Schlang	Dec 1985	A
4656665	Pennebaker	Apr 1987	A
4836026	P'an et al.	Jun 1989	A
4903312	Sato	Feb 1990	A
4992863	Moriya	Feb 1991	A
5020112	Chou	May 1991	A
5063604	Weiman	Nov 1991	A
5101448	Kawachiya et al.	Mar 1992	A
5124810	Seto	Jun 1992	A
5151260	Contursi et al.	Sep 1992	A
5159667	Borrey et al.	Oct 1992	A
5181260	Kurosu et al.	Jan 1993	A
5202934	Miyakawa et al.	Apr 1993	A
5220621	Saitoh	Jun 1993	A
5268967	Jang et al.	Dec 1993	A
5282055	Suzuki	Jan 1994	A
5293429	Pizano et al.	Mar 1994	A
5313527	Guberman et al.	May 1994	A
5317646	Sang, Jr. et al.	May 1994	A
5344132	LeBrun et al.	Sep 1994	A
5353673	Lynch	Oct 1994	A
5355547	Fitjer	Oct 1994	A
5375197	Kang	Dec 1994	A
5430810	Saeki	Jul 1995	A
5467407	Guberman et al.	Nov 1995	A
5473742	Polyakov et al.	Dec 1995	A
5546474	Zuniga	Aug 1996	A
5563723	Beaulieu et al.	Oct 1996	A
5563966	Ise et al.	Oct 1996	A
5586199	Kanda et al.	Dec 1996	A
5594815	Fast et al.	Jan 1997	A
5596655	Lopez	Jan 1997	A
5602964	Barrett	Feb 1997	A
5629989	Osada	May 1997	A
5652663	Zelten	Jul 1997	A
5668890	Winkelman	Sep 1997	A
5680525	Sakai et al.	Oct 1997	A
5696611	Nishimura et al.	Dec 1997	A
5696805	Gaborski et al.	Dec 1997	A
5699244	Clark, Jr. et al.	Dec 1997	A
5717794	Koga et al.	Feb 1998	A
5721940	Luther et al.	Feb 1998	A
5757963	Ozaki et al.	May 1998	A
5760912	Itoh	Jun 1998	A
5781665	Cullen et al.	Jul 1998	A
5818978	Al-Hussein	Oct 1998	A
5822454	Rangarajan	Oct 1998	A
5825915	Michimoto et al.	Oct 1998	A
5832138	Nakanishi et al.	Nov 1998	A
5839019	Ito	Nov 1998	A
5848184	Taylor et al.	Dec 1998	A
5857029	Patel	Jan 1999	A
5867264	Hinnrichs	Feb 1999	A
5899978	Irwin	May 1999	A
5937084	Crabtree et al.	Aug 1999	A
5953388	Walnut et al.	Sep 1999	A
5987172	Michael	Nov 1999	A
6002489	Murai et al.	Dec 1999	A
6005958	Farmer et al.	Dec 1999	A
6005968	Granger	Dec 1999	A
6009191	Julier	Dec 1999	A
6009196	Mahoney	Dec 1999	A
6011595	Henderson et al.	Jan 2000	A
6016361	Hongu et al.	Jan 2000	A
6038348	Carley	Mar 2000	A
6055968	Sasaki et al.	May 2000	A
6067385	Cullen et al.	May 2000	A
6072916	Suzuki	Jun 2000	A
6073148	Rowe et al.	Jun 2000	A
6098065	Skillen et al.	Aug 2000	A
6104830	Schistad	Aug 2000	A
6104840	Ejiri et al.	Aug 2000	A
6118544	Rao	Sep 2000	A
6118552	Suzuki et al.	Sep 2000	A
6154217	Aldrich	Nov 2000	A
6192360	Dumais et al.	Feb 2001	B1
6215469	Mori et al.	Apr 2001	B1
6219158	Dawe	Apr 2001	B1
6219773	Garibay, Jr. et al.	Apr 2001	B1
6223223	Kumpf et al.	Apr 2001	B1
6229625	Nakatsuka	May 2001	B1
6233059	Kodaira et al.	May 2001	B1
6263122	Simske et al.	Jul 2001	B1
6292168	Venable et al.	Sep 2001	B1
6327581	Platt	Dec 2001	B1
6337925	Cohen et al.	Jan 2002	B1
6347152	Shinagawa et al.	Feb 2002	B1
6347162	Suzuki	Feb 2002	B1
6356647	Bober et al.	Mar 2002	B1
6370277	Borrey et al.	Apr 2002	B1
6385346	Gillihan et al.	May 2002	B1
6393147	Danneels et al.	May 2002	B2
6396599	Patton et al.	May 2002	B1
6408094	Mirzaoff et al.	Jun 2002	B1
6408105	Maruo	Jun 2002	B1
6424742	Yamamoto et al.	Jul 2002	B2
6433896	Ueda et al.	Aug 2002	B1
6456738	Tsukasa	Sep 2002	B1
6463430	Brady et al.	Oct 2002	B1
6469801	Telle	Oct 2002	B1
6473198	Matama	Oct 2002	B1
6473535	Takaoka	Oct 2002	B1
6480304	Os et al.	Nov 2002	B1
6480624	Horie et al.	Nov 2002	B1
6501855	Zelinski	Dec 2002	B1
6512848	Wang et al.	Jan 2003	B2
6522791	Nagarajan	Feb 2003	B2
6525840	Haraguchi et al.	Feb 2003	B1
6563531	Matama	May 2003	B1
6601026	Appelt et al.	Jul 2003	B2
6614930	Agnihotri et al.	Sep 2003	B1
6621595	Fan et al.	Sep 2003	B1
6628416	Hsu et al.	Sep 2003	B1
6628808	Bach et al.	Sep 2003	B1
6633857	Tipping	Oct 2003	B1
6643413	Shum et al.	Nov 2003	B1
6646765	Barker et al.	Nov 2003	B1
6658147	Gorbatov et al.	Dec 2003	B2
6665425	Sampath et al.	Dec 2003	B1
6667774	Berman et al.	Dec 2003	B2
6675159	Lin et al.	Jan 2004	B1
6701009	Makoto et al.	Mar 2004	B1
6704441	Inagaki et al.	Mar 2004	B1
6724916	Shyu	Apr 2004	B1
6732046	Joshi	May 2004	B1
6748109	Yamaguchi	Jun 2004	B1
6751349	Matama	Jun 2004	B2
6757081	Fan et al.	Jun 2004	B1
6757427	Hongu	Jun 2004	B1
6763515	Vazquez et al.	Jul 2004	B1
6765685	Yu	Jul 2004	B1
6778684	Bollman	Aug 2004	B1
6781375	Miyazaki et al.	Aug 2004	B2
6788830	Morikawa	Sep 2004	B1
6789069	Barnhill et al.	Sep 2004	B1
6801658	Morita et al.	Oct 2004	B2
6826311	Wilt	Nov 2004	B2
6831755	Narushima et al.	Dec 2004	B1
6839466	Venable	Jan 2005	B2
6850653	Abe	Feb 2005	B2
6873721	Beyerer et al.	Mar 2005	B1
6882983	Furphy et al.	Apr 2005	B2
6898601	Amado et al.	May 2005	B2
6901170	Terada et al.	May 2005	B1
6917438	Yoda et al.	Jul 2005	B1
6917709	Zelinski	Jul 2005	B2
6921220	Aiyama	Jul 2005	B2
6950555	Filatov et al.	Sep 2005	B2
6987534	Seta	Jan 2006	B1
6989914	Iwaki	Jan 2006	B2
7006707	Peterson	Feb 2006	B2
7016549	Utagawa	Mar 2006	B1
7017108	Wan	Mar 2006	B1
7020320	Filatov	Mar 2006	B2
7023447	Luo et al.	Apr 2006	B2
7027181	Takamori	Apr 2006	B2
7038713	Matama	May 2006	B1
7042603	Masao et al.	May 2006	B2
7043080	Dolan	May 2006	B1
7054036	Hirayama	May 2006	B2
7081975	Yoda et al.	Jul 2006	B2
7082426	Musgrove et al.	Jul 2006	B2
7123387	Cheng et al.	Oct 2006	B2
7130471	Bossut et al.	Oct 2006	B2
7145699	Dolan	Dec 2006	B2
7149347	Wnek	Dec 2006	B1
7167281	Fujimoto et al.	Jan 2007	B1
7168614	Kotovich et al.	Jan 2007	B2
7173732	Matama	Feb 2007	B2
7174043	Lossev et al.	Feb 2007	B2
7177049	Karidi	Feb 2007	B2
7181082	Feng	Feb 2007	B2
7184929	Goodman	Feb 2007	B2
7194471	Nagatsuka et al.	Mar 2007	B1
7197158	Camara et al.	Mar 2007	B2
7201323	Kotovich et al.	Apr 2007	B2
7209599	Simske et al.	Apr 2007	B2
7228314	Kawamoto et al.	Jun 2007	B2
7249717	Kotovich et al.	Jul 2007	B2
7251777	Valtchev et al.	Jul 2007	B1
7253836	Suzuki et al.	Aug 2007	B1
7263221	Moriwaki	Aug 2007	B1
7266768	Ferlitsch et al.	Sep 2007	B2
7286177	Cooper	Oct 2007	B2
7298897	Dominguez et al.	Nov 2007	B1
7317828	Suzuki et al.	Jan 2008	B2
7337389	Woolf et al.	Feb 2008	B1
7339585	Verstraelen et al.	Mar 2008	B2
7340376	Goodman	Mar 2008	B2
7349888	Heidenreich et al.	Mar 2008	B1
7365881	Burns et al.	Apr 2008	B2
7366705	Zeng et al.	Apr 2008	B2
7382921	Lossev et al.	Jun 2008	B2
7386527	Harris et al.	Jun 2008	B2
7392426	Wolfe et al.	Jun 2008	B2
7403008	Blank et al.	Jul 2008	B2
7403313	Kuo	Jul 2008	B2
7406183	Emerson et al.	Jul 2008	B2
7409092	Srinivasa	Aug 2008	B2
7409633	Lerner et al.	Aug 2008	B2
7416131	Fortune et al.	Aug 2008	B2
7426293	Chien et al.	Sep 2008	B2
7430059	Rodrigues et al.	Sep 2008	B2
7430066	Hsu et al.	Sep 2008	B2
7430310	Kotovich et al.	Sep 2008	B2
7447377	Takahira	Nov 2008	B2
7464066	Zelinski et al.	Dec 2008	B2
7478332	Buttner et al.	Jan 2009	B2
7487438	Withers	Feb 2009	B1
7492478	Une	Feb 2009	B2
7492943	Li et al.	Feb 2009	B2
7515313	Cheng	Apr 2009	B2
7515772	Li et al.	Apr 2009	B2
7528883	Hsu	May 2009	B2
7545529	Borrey et al.	Jun 2009	B2
7553095	Kimura	Jun 2009	B2
7562060	Sindhwani et al.	Jul 2009	B2
7580557	Zavadsky et al.	Aug 2009	B2
7639387	Hull et al.	Dec 2009	B2
7643665	Zavadsky et al.	Jan 2010	B2
7651286	Tischler	Jan 2010	B2
7655685	McElroy et al.	Feb 2010	B2
7657091	Postnikov et al.	Feb 2010	B2
7665061	Kothari et al.	Feb 2010	B2
7673799	Hart et al.	Mar 2010	B2
7702162	Cheong et al.	Apr 2010	B2
7735721	Ma et al.	Jun 2010	B1
7738730	Hawley	Jun 2010	B2
7739127	Hall	Jun 2010	B1
7761391	Schmidtler et al.	Jul 2010	B2
7778457	Nepomniachtchi et al.	Aug 2010	B2
7782384	Kelly	Aug 2010	B2
7787695	Nepomniachtchi et al.	Aug 2010	B2
7937345	Schmidtler et al.	May 2011	B2
7949167	Krishnan et al.	May 2011	B2
7949176	Nepomniachtchi	May 2011	B2
7949660	Green et al.	May 2011	B2
7953268	Nepomniachtchi	May 2011	B2
7958067	Schmidtler et al.	Jun 2011	B2
7978900	Nepomniachtchi et al.	Jul 2011	B2
7999961	Wanda	Aug 2011	B2
8000514	Nepomniachtchi et al.	Aug 2011	B2
8064710	Mizoguchi	Nov 2011	B2
8073263	Hull et al.	Dec 2011	B2
8078958	Cottrille et al.	Dec 2011	B2
8081227	Kim et al.	Dec 2011	B1
8094976	Berard et al.	Jan 2012	B2
8135656	Evanitsky	Mar 2012	B2
8184156	Mino et al.	May 2012	B2
8194965	Lossev et al.	Jun 2012	B2
8213687	Fan	Jul 2012	B2
8239335	Schmidtler et al.	Aug 2012	B2
8244031	Cho et al.	Aug 2012	B2
8279465	Couchman	Oct 2012	B2
8295599	Katougi et al.	Oct 2012	B2
8311296	Filatov et al.	Nov 2012	B2
8326015	Nepomniachtchi	Dec 2012	B2
8345981	Schmidtler et al.	Jan 2013	B2
8354981	Kawasaki et al.	Jan 2013	B2
8374977	Schmidtler et al.	Feb 2013	B2
8379914	Nepomniachtchi et al.	Feb 2013	B2
8385647	Hawley et al.	Feb 2013	B2
8406480	Grigsby et al.	Mar 2013	B2
8441548	Nechyba et al.	May 2013	B1
8452098	Nepomniachtchi et al.	May 2013	B2
8483473	Roach et al.	Jul 2013	B2
8503797	Turkelson et al.	Aug 2013	B2
8515163	Cho et al.	Aug 2013	B2
8515208	Minerich	Aug 2013	B2
8526739	Schmidtler et al.	Sep 2013	B2
8532419	Coleman	Sep 2013	B2
8559766	Tilt et al.	Oct 2013	B2
8577118	Nepomniachtchi et al.	Nov 2013	B2
8582862	Nepomniachtchi et al.	Nov 2013	B2
8587818	Imaizumi et al.	Nov 2013	B2
8620058	Nepomniachtchi et al.	Dec 2013	B2
8639621	Ellis et al.	Jan 2014	B1
8677249	Buttner et al.	Mar 2014	B2
8693043	Schmidtler et al.	Apr 2014	B2
8705836	Gorski et al.	Apr 2014	B2
8719197	Schmidtler et al.	May 2014	B2
8749839	Borrey et al.	Jun 2014	B2
8774516	Amtrup et al.	Jul 2014	B2
8823991	Borrey et al.	Sep 2014	B2
8855375	Macciola et al.	Oct 2014	B2
8855425	Schmidtler et al.	Oct 2014	B2
8879120	Thrasher et al.	Nov 2014	B2
8885229	Amtrup et al.	Nov 2014	B1
8955743	Block et al.	Feb 2015	B1
8971587	Macciola et al.	Mar 2015	B2
8989515	Shustorovich et al.	Mar 2015	B2
9058515	Amtrup et al.	Jun 2015	B1
9117117	Macciola et al.	Aug 2015	B2
9129210	Borrey et al.	Sep 2015	B2
9137417	Macciola et al.	Sep 2015	B2
9158967	Shustorovich et al.	Oct 2015	B2
9165187	Macciola et al.	Oct 2015	B2
9165188	Thrasher et al.	Oct 2015	B2
9275281	Macciola	Mar 2016	B2
9311531	Amtrup et al.	Apr 2016	B2
9342742	Amtrup et al.	May 2016	B2
9355312	Amtrup et al.	May 2016	B2
9514357	Macciola et al.	Dec 2016	B2
20010027420	Boublik et al.	Oct 2001	A1
20020030831	Kinjo	Mar 2002	A1
20020054693	Elmenhurst	May 2002	A1
20020069218	Sull et al.	Jun 2002	A1
20020113801	Reavy et al.	Aug 2002	A1
20020122071	Camara et al.	Sep 2002	A1
20020126313	Namizuka	Sep 2002	A1
20020165717	Solmer et al.	Nov 2002	A1
20030002068	Constantin et al.	Jan 2003	A1
20030026479	Thomas et al.	Feb 2003	A1
20030030638	Astrom et al.	Feb 2003	A1
20030044012	Eden	Mar 2003	A1
20030046445	Witt et al.	Mar 2003	A1
20030053696	Schmidt et al.	Mar 2003	A1
20030063213	Poplin	Apr 2003	A1
20030086615	Dance	May 2003	A1
20030095709	Zhou	May 2003	A1
20030101161	Ferguson et al.	May 2003	A1
20030117511	Belz et al.	Jun 2003	A1
20030120653	Brady et al.	Jun 2003	A1
20030142328	McDaniel et al.	Jul 2003	A1
20030151674	Lin	Aug 2003	A1
20030156201	Zhang	Aug 2003	A1
20030197063	Longacre	Oct 2003	A1
20030210428	Bevlin et al.	Nov 2003	A1
20030223615	Keaton et al.	Dec 2003	A1
20040021909	Kikuoka	Feb 2004	A1
20040022437	Beardsley	Feb 2004	A1
20040083119	Schunder et al.	Apr 2004	A1
20040090458	Yu et al.	May 2004	A1
20040093119	Gunnarsson et al.	May 2004	A1
20040102989	Jang et al.	May 2004	A1
20040111453	Harris et al.	Jun 2004	A1
20040143547	Mersky	Jul 2004	A1
20040143796	Lerner et al.	Jul 2004	A1
20040169873	Nagarajan	Sep 2004	A1
20040169889	Sawada	Sep 2004	A1
20040175033	Matama	Sep 2004	A1
20040181482	Yap	Sep 2004	A1
20040190019	Li et al.	Sep 2004	A1
20040261084	Rosenbloom et al.	Dec 2004	A1
20040263639	Sadovsky et al.	Dec 2004	A1
20050021360	Miller et al.	Jan 2005	A1
20050030602	Gregson et al.	Feb 2005	A1
20050046887	Shibata et al.	Mar 2005	A1
20050050060	Damm et al.	Mar 2005	A1
20050054342	Otsuka	Mar 2005	A1
20050060162	Mohit et al.	Mar 2005	A1
20050063585	Matsuura	Mar 2005	A1
20050065903	Zhang et al.	Mar 2005	A1
20050080844	Dathathraya et al.	Apr 2005	A1
20050100209	Lewis et al.	May 2005	A1
20050131780	Princen	Jun 2005	A1
20050134935	Schmidtler et al.	Jun 2005	A1
20050141777	Kuwata	Jun 2005	A1
20050151990	Ishikawa et al.	Jul 2005	A1
20050160065	Seeman	Jul 2005	A1
20050180628	Curry et al.	Aug 2005	A1
20050180632	Aradhye et al.	Aug 2005	A1
20050193325	Epstein	Sep 2005	A1
20050204058	Philbrick et al.	Sep 2005	A1
20050206753	Sakurai et al.	Sep 2005	A1
20050212925	Lefebure et al.	Sep 2005	A1
20050216426	Weston et al.	Sep 2005	A1
20050228591	Hur et al.	Oct 2005	A1
20050234955	Zeng et al.	Oct 2005	A1
20050246262	Aggarwal et al.	Nov 2005	A1
20050265618	Jebara	Dec 2005	A1
20050271265	Wang et al.	Dec 2005	A1
20060013463	Ramsay et al.	Jan 2006	A1
20060023271	Boay et al.	Feb 2006	A1
20060031344	Mishima et al.	Feb 2006	A1
20060048046	Joshi et al.	Mar 2006	A1
20060074821	Cristianini	Apr 2006	A1
20060089907	Kohlmaier et al.	Apr 2006	A1
20060093208	Li et al.	May 2006	A1
20060095373	Venkatasubramanian et al.	May 2006	A1
20060095374	Lo et al.	May 2006	A1
20060098899	King et al.	May 2006	A1
20060112340	Mohr et al.	May 2006	A1
20060114488	Motamed	Jun 2006	A1
20060115153	Bhattacharjya	Jun 2006	A1
20060120609	Ivanov et al.	Jun 2006	A1
20060147113	Han	Jul 2006	A1
20060159364	Poon et al.	Jul 2006	A1
20060161646	Chene et al.	Jul 2006	A1
20060195491	Nieland et al.	Aug 2006	A1
20060203107	Steinberg et al.	Sep 2006	A1
20060206628	Erez	Sep 2006	A1
20060212413	Rujan et al.	Sep 2006	A1
20060215231	Borrey et al.	Sep 2006	A1
20060219773	Richardson	Oct 2006	A1
20060222239	Bargeron et al.	Oct 2006	A1
20060235732	Miller et al.	Oct 2006	A1
20060235812	Rifkin et al.	Oct 2006	A1
20060236304	Luo et al.	Oct 2006	A1
20060242180	Graf et al.	Oct 2006	A1
20060256371	King et al.	Nov 2006	A1
20060256392	Van Hoof et al.	Nov 2006	A1
20060262962	Hull et al.	Nov 2006	A1
20060263134	Beppu	Nov 2006	A1
20060265640	Albornoz et al.	Nov 2006	A1
20060268352	Tanigawa et al.	Nov 2006	A1
20060268356	Shih et al.	Nov 2006	A1
20060268369	Kuo	Nov 2006	A1
20060279798	Rudolph et al.	Dec 2006	A1
20060282442	Lennon et al.	Dec 2006	A1
20060282463	Rudolph et al.	Dec 2006	A1
20060282762	Diamond et al.	Dec 2006	A1
20060288015	Schirripa et al.	Dec 2006	A1
20060294154	Shimizu	Dec 2006	A1
20070002348	Hagiwara	Jan 2007	A1
20070003155	Miller et al.	Jan 2007	A1
20070005341	Burges et al.	Jan 2007	A1
20070016848	Rosenoff et al.	Jan 2007	A1
20070030540	Cheng et al.	Feb 2007	A1
20070035780	Kanno	Feb 2007	A1
20070046957	Jacobs et al.	Mar 2007	A1
20070046982	Hull et al.	Mar 2007	A1
20070047782	Hull et al.	Mar 2007	A1
20070065033	Hernandez et al.	Mar 2007	A1
20070086667	Dai et al.	Apr 2007	A1
20070109590	Hagiwara	May 2007	A1
20070118794	Hollander et al.	May 2007	A1
20070128899	Mayer	Jun 2007	A1
20070133862	Gold et al.	Jun 2007	A1
20070165801	Devolites et al.	Jul 2007	A1
20070177818	Teshima et al.	Aug 2007	A1
20070239642	Sindhwani et al.	Oct 2007	A1
20070250416	Beach et al.	Oct 2007	A1
20070252907	Hsu	Nov 2007	A1
20070260588	Biazetti et al.	Nov 2007	A1
20080005678	Buttner et al.	Jan 2008	A1
20080082352	Schmidtler et al.	Apr 2008	A1
20080086432	Schmidtler et al.	Apr 2008	A1
20080086433	Schmidtler et al.	Apr 2008	A1
20080097936	Schmidtler et al.	Apr 2008	A1
20080130992	Fujii	Jun 2008	A1
20080133388	Alekseev et al.	Jun 2008	A1
20080137971	King et al.	Jun 2008	A1
20080144881	Fortune et al.	Jun 2008	A1
20080147561	Euchner et al.	Jun 2008	A1
20080147790	Malaney et al.	Jun 2008	A1
20080175476	Ohk et al.	Jul 2008	A1
20080177643	Matthews et al.	Jul 2008	A1
20080183576	Kim et al.	Jul 2008	A1
20080199081	Kimura et al.	Aug 2008	A1
20080212115	Konishi	Sep 2008	A1
20080215489	Lawson et al.	Sep 2008	A1
20080219543	Csulits et al.	Sep 2008	A1
20080235766	Wallos et al.	Sep 2008	A1
20080253647	Cho et al.	Oct 2008	A1
20080294737	Kim	Nov 2008	A1
20090015687	Shinkai et al.	Jan 2009	A1
20090089078	Bursey	Apr 2009	A1
20090103808	Dey et al.	Apr 2009	A1
20090141985	Sheinin et al.	Jun 2009	A1
20090175537	Tribelhorn et al.	Jul 2009	A1
20090185241	Nepomniachtchi	Jul 2009	A1
20090214112	Borrey et al.	Aug 2009	A1
20090228499	Schmidtler et al.	Sep 2009	A1
20090327250	Green et al.	Dec 2009	A1
20100060910	Fechter	Mar 2010	A1
20100060915	Suzuki et al.	Mar 2010	A1
20100166318	Ben-Horesh et al.	Jul 2010	A1
20100169250	Schmidtler et al.	Jul 2010	A1
20100202698	Schmidtler et al.	Aug 2010	A1
20100202701	Basri et al.	Aug 2010	A1
20100209006	Grigsby et al.	Aug 2010	A1
20100232706	Forutanpour	Sep 2010	A1
20100280859	Frederick, II	Nov 2010	A1
20100331043	Chapman et al.	Dec 2010	A1
20110013039	Aisaka et al.	Jan 2011	A1
20110025860	Katougi et al.	Feb 2011	A1
20110032570	Imaizumi et al.	Feb 2011	A1
20110035284	Moshfeghi	Feb 2011	A1
20110091092	Nepomniachtchi et al.	Apr 2011	A1
20110116716	Kwon et al.	May 2011	A1
20110145178	Schmidtler et al.	Jun 2011	A1
20110182500	Esposito et al.	Jul 2011	A1
20110194127	Nagakoshi et al.	Aug 2011	A1
20110196870	Schmidtler et al.	Aug 2011	A1
20110246076	Su et al.	Oct 2011	A1
20110249905	Singh et al.	Oct 2011	A1
20110280450	Nepomniachtchi et al.	Nov 2011	A1
20110285873	Showering	Nov 2011	A1
20110285874	Showering et al.	Nov 2011	A1
20120038549	Mandella et al.	Feb 2012	A1
20120057756	Yoon et al.	Mar 2012	A1
20120075442	Vujic	Mar 2012	A1
20120113489	Heit et al.	May 2012	A1
20120114249	Conwell	May 2012	A1
20120116957	Zanzot et al.	May 2012	A1
20120162527	Baker	Jun 2012	A1
20120194692	Mers et al.	Aug 2012	A1
20120230577	Calman et al.	Sep 2012	A1
20120272192	Grossman et al.	Oct 2012	A1
20120284122	Brandis	Nov 2012	A1
20130004076	Koo	Jan 2013	A1
20130022231	Nepomniachtchi et al.	Jan 2013	A1
20130027757	Lee et al.	Jan 2013	A1
20130060596	Gu et al.	Mar 2013	A1
20130073459	Zacarias et al.	Mar 2013	A1
20130080347	Paul et al.	Mar 2013	A1
20130088757	Schmidtler et al.	Apr 2013	A1
20130090969	Rivere	Apr 2013	A1
20130117175	Hanson	May 2013	A1
20130121610	Chen et al.	May 2013	A1
20130124414	Roach et al.	May 2013	A1
20130152176	Courtney et al.	Jun 2013	A1
20130182002	Macciola et al.	Jul 2013	A1
20130182105	Fahn	Jul 2013	A1
20130182128	Amtrup et al.	Jul 2013	A1
20130182292	Thrasher et al.	Jul 2013	A1
20130182951	Shustorovich et al.	Jul 2013	A1
20130182959	Thrasher et al.	Jul 2013	A1
20130182970	Shustorovich et al.	Jul 2013	A1
20130182973	Macciola et al.	Jul 2013	A1
20130185618	Macciola et al.	Jul 2013	A1
20130188865	Saha et al.	Jul 2013	A1
20130223762	Nagamasa	Aug 2013	A1
20130230246	Nuggehalli	Sep 2013	A1
20130251280	Borrey et al.	Sep 2013	A1
20130268378	Yovin	Oct 2013	A1
20130268430	Lopez et al.	Oct 2013	A1
20130287265	Nepomniachtchi et al.	Oct 2013	A1
20130287284	Nepomniachtchi et al.	Oct 2013	A1
20130290036	Strange	Oct 2013	A1
20130297353	Strange et al.	Nov 2013	A1
20130308832	Schmidtler et al.	Nov 2013	A1
20140003721	Saund	Jan 2014	A1
20140006129	Heath	Jan 2014	A1
20140006198	Daly et al.	Jan 2014	A1
20140012754	Hanson et al.	Jan 2014	A1
20140079294	Amtrup et al.	Mar 2014	A1
20140108456	Ramachandrula et al.	Apr 2014	A1
20140153787	Schmidtler et al.	Jun 2014	A1
20140153830	Amtrup et al.	Jun 2014	A1
20140164914	Schmidtler et al.	Jun 2014	A1
20140172687	Chirehdast	Jun 2014	A1
20140201612	Buttner et al.	Jul 2014	A1
20140207717	Schmidtler et al.	Jul 2014	A1
20140233068	Borrey et al.	Aug 2014	A1
20140254887	Amtrup et al.	Sep 2014	A1
20140270349	Amtrup et al.	Sep 2014	A1
20140270439	Chen	Sep 2014	A1
20140270536	Amtrup et al.	Sep 2014	A1
20140328520	Macciola et al.	Nov 2014	A1
20140376060	Bocharov et al.	Dec 2014	A1
20150120564	Smith et al.	Apr 2015	A1
20150161765	Kota et al.	Jun 2015	A1
20160028921	Thrasher et al.	Jan 2016	A1
20160125613	Shustorovich et al.	May 2016	A1
20160217319	Bhanu et al.	Jul 2016	A1
20170046788	Macciola et al.	Feb 2017	A1
20170147572	Kilby et al.	May 2017	A1

Foreign Referenced Citations (53)

Number	Date	Country
101052991	Oct 2007	CN
0549329	Jun 1993	EP
0767578	Apr 1997	EP
0936804	Aug 1999	EP
1054331	Nov 2000	EP
1128659	Aug 2001	EP
1317133	Jun 2003	EP
1319133	Jun 2003	EP
1422920	May 2004	EP
1956518	Aug 2008	EP
1959363	Aug 2008	EP
1976259	Oct 2008	EP
2107480	Oct 2009	EP
H07260701	Oct 1995	JP
H0962826	Mar 1997	JP
H09116720	May 1997	JP
H11118444	Apr 1999	JP
2000067065	Mar 2000	JP
2000103628	Apr 2000	JP
2000354144	Dec 2000	JP
2001309128	Nov 2001	JP
2002024258	Jan 2002	JP
2002519766	Jul 2002	JP
2002312385	Oct 2002	JP
2003091521	Mar 2003	JP
2003196357	Jul 2003	JP
2003234888	Aug 2003	JP
2003303315	Oct 2003	JP
2005018678	Jan 2005	JP
2005085173	Mar 2005	JP
2005173730	Jun 2005	JP
2006031379	Feb 2006	JP
2006209588	Aug 2006	JP
2006330863	Dec 2006	JP
2008134683	Jun 2008	JP
2011034387	Feb 2011	JP
2011055467	Mar 2011	JP
2012009033	Jan 2012	JP
2012517637	Aug 2012	JP
2013196357	Sep 2013	JP
5462286	Apr 2014	JP
401553	Aug 2000	TW
9604749	Feb 1996	WO
9847098	Oct 1998	WO
9967731	Dec 1999	WO
0263812	Aug 2002	WO
02063812	Aug 2002	WO
2004053630	Jun 2004	WO
2004056360	Jul 2004	WO
2006104627	Oct 2006	WO
2008008142	Jan 2008	WO
2010030056	Mar 2010	WO
2010096192	Aug 2010	WO

Non-Patent Literature Citations (59)

Entry
Non-Final Office Action from U.S Appl. No. 14/927,359, dated Nov. 21, 2016.
Final Office Action from U.S Appl. No. 14/814,455, dated Dec. 16, 2016.
International Search Report and Written Opinion from International Application No. PCT/US14/26569, dated Aug. 12, 2014.
Gllavata et al., “Finding Text in Images Via Local Thresholding,” International Symposium on Signal Processing and Information Technology, Dec. 2003, pp. 539-542.
Zunino et al., “Vector Quantization for License-Plate Location and Image Coding,” IEEE Transactions on Industrial Electronics, vol. 47, Issue 1, Feb. 2000, pp. 159-167.
Extended European Search Report from European Application No. 14775259.6, dated Jun. 1, 2016.
Non-Final Office Action from U.S. Appl. No. 13/898,407, dated Aug. 1, 2013.
Final Office Action from U.S. Appl. No. 13/898,407, dated Jan. 13, 2014.
Notice of Allowance from U.S. Appl. No. 13/898,407, dated Apr. 23, 2014.
Non-Final Office Action from U.S. Appl. No. 14/340,460, dated Jan. 16, 2015.
Notice of Allowance from U.S. Appl. No. 14/340,460, dated Apr. 28, 2015.
Office Action from Japanese Patent Application No. 2014-552356, dated Jun. 2, 2015.
Office Action from Taiwan Application No. 102101177, dated Dec. 17, 2014.
Notice of Allowance from U.S. Appl. No. 14/220,023, dated Jan. 30, 2015.
Notice of Allowance from U.S. Appl. No. 14/220,029, dated Feb. 11, 2015.
International Search Report and Written Opinion from International Application No. PCT/US2013/021336, dated May 23, 2013.
Non-Final Office Action from U.S. Appl. No. 13/740,127, dated Oct. 27, 2014.
Notice of Allowance from U.S. Appl. No. 13/740,131, dated Oct. 27, 2014.
Final Office Action from U.S. Appl. No. 13/740,134, dated Mar. 3, 2015.
Non-Final Office Action from U.S. Appl. No. 13/740,134, dated Oct. 10, 2014.
Non-Final Office Action from U.S. Appl. No. 13/740,138, dated Dec. 1, 2014.
Notice of Allowance from U.S. Appl. No. 13/740,139, dated Aug. 29, 2014.
Notice of Allowance from U.S. Appl. No. 13/740,145, dated Mar. 30, 2015.
Non-Final Office Action from U.S. Appl. No. 13/740,145, dated Sep. 29, 2014.
Notice of Allowance from Taiwan Patent Application No. 102101177, dated Apr. 24, 2015.
Notice of Allowance from U.S. Appl. No. 13/740,138, dated Jun. 5, 2015.
Notice of Allowance from U.S. Appl. No. 13/740,127, dated Jun. 8, 2015.
Notice of Allowance from U.S. Appl. No. 14/569,375, dated Apr. 15, 2015.
Notice of Allowance from U.S. Appl. No. 13/740,134, dated May 29, 2015.
Notice of Allowability from U.S. Appl. No. 13/740,145, dated May 26, 2015.
Corrected Notice of Allowability from U.S. Appl. No. 13/740,138, dated Jul. 8, 2018.
Non-Final Office Action from U.S. Appl. No. 13/740,127, dated Feb. 23, 2015.
Final Office Action from U.S. Appl. No. 13/740,134, dated Mar. 3, 3015.
Notice of Allowance from U.S. Appl. No. 14/804,276, dated Oct. 21, 2015.
Extended Europrean Search Report from European Application No. 13738301.4, dated Nov. 17, 2015.
Notice of Allowance from U.S. Appl. No. 13/740,145, dated Jan. 15, 2016.
Office Action from Taiwan Patent Application No. 102101177, dated Dec. 17, 2014.
Non-Final Office Action from U.S. Appl. No. 13/740,141, dated Oct. 16, 2015.
Notice of Allowance from U.S. Appl. No. 13/740,145, dated Sep. 8, 2015.
Notice of Allowance from U.S. Appl. No. 14/334,558, dated Sep. 10, 2014.
Notice of Allowance from U.S. Appl. No. 13/740,123, dated Jul. 10, 2014.
Intsig Information Co., Ltd., “CamScanner,” www.intsig.com/en/camscanner.html, retrieved Oct. 25, 2012.
Intsig Information Co., Ltd., “Product Descriptions,” www.intsig.com/en/product.html, retrieved Oct. 25, 2012.
Non-Final Office Action from U.S. Appl. No. 14/814,455, dated Jun. 17, 2016.
Final Office Action from U.S. Appl. No. 13/740,141, dated May 5, 2016.
Thrasher, C. W. et al., U.S. Appl. No. 15/214,351, filed Jul. 19, 2016.
Notice of Allowance from U.S. Appl. No. 13/740,141, dated Jul. 29, 2016.
International Search Report and Written Opinion from International Application No. PCT/US2016/043207, dated Oct. 21, 2016.
Office Action from Chinese Patent Application No. 201380004057.2, dated Feb. 27, 2017.
Notice of Allowance from U.S. Appl. No. 14/814,455, dated Mar. 30, 2017.
Non-Final Office Action from U.S. Appl. No. 14/932,902, dated Apr. 11, 2017.
Non-Final Office Action from U.S. Appl. No. 15/390,321, dated Mar. 17, 2017.
Notice of Allowance from U.S. Appl. No. 15/146,848, dated Apr. 13, 2017.
Notice of Allowance from U.S. Appl. No. 15/191,442, dated Apr. 24, 2017.
Final Office Action from U.S. Appl. No. 14/927,359, dated Apr. 28, 2017.
Notice of Allowance from U.S. Appl. No. 15/234,969, dated May 8, 2017.
Notice of Allowance from U.S. Appl. No. 14/814,455, dated May 26, 2017.
Corrected Notice of Allowance from U.S. Appl. No. 15/191,442, dated May 26, 2017.
Extended European Search Report from European Application No. 14881675.4, dated Jun. 7, 2017.

Related Publications (1)

	Number	Date	Country
	20150339526 A1	Nov 2015	US

Provisional Applications (1)

	Number	Date	Country
	61780747	Mar 2013	US

Continuations (1)

	Number	Date	Country
Parent	14209825	Mar 2014	US
Child	14818196		US

Systems and methods for classifying objects in digital images captured using mobile devices

Information

Patent Number

Date Filed

Date Issued

Inventors

Original Assignees

Examiners

Agents

CPC

Field of Search

CPC

International Classifications

Abstract