Using gestalt information to identify locations in printed information

Abstract

A facility for identifying a location in a printed document is described. The facility obtains an image of the printed document, and extracts gestalt information from text occurring in the image of the printed document. The facility compares the extracted gestalt information to an index of documents and, based upon this comparison, identifies a document that includes the gestalt information.

Description

BACKGROUND

Current image recognition, such as techniques from recognizing text, images etc., may suffer when lighting is poor, the image is of low quality, the user is unskilled, and so on.

BRIEF DESCRIPTION OF THE DRAWINGS

FIGS. 1-6 are document diagrams showing many different types of information derived from the spaces between words.

FIG. 7 and the following discussion provide a brief, general description of a suitable environment in which the system can be implemented.

DETAILED DESCRIPTION

A software and/or hardware facility (“the facility”) is described that uses gestalt information to recognize a place or location in a document or other display of information. This can be mapped to an index of text, or an index of both text and text layout information.

In some embodiments, the facility determines a position in a text- or character-based document or other display of information. In some cases, the document includes words, such as a western language newspaper, book, magazine, and so on. In some cases, the document includes characters, such as an eastern language newspaper, book, magazine, and so on. In some cases, the document includes images and text, such as an advertisement.

The facility receives gestalt information associated with a location in a document (such as by taking an image of the document using a camera or other imaging component), compares the gestalt information to an index associated with the document or associated with a group of documents, and identifies the location in the document based on the comparison.

In some examples, the facility may extract information associated with spaces between words printed on a document. The information may be distances between spaces (horizontally, vertically, orthogonally, and so on), the width of the spaces, and so on.

In some examples, the facility may extract information associated with densities of characters printed on a document. The information may include relative densities between characters (horizontally, vertically, orthogonally, and so on), relative densities between grouped pairs of characters, absolute density information, and so on.

Various examples of the facility will now be described. The following description provides specific details for a thorough understanding and enabling description of these examples. One skilled in the relevant art will understand, however, that the facility may be practiced without many of these details. Likewise, one skilled in the relevant art will also understand that the facility can include many other obvious features not described in detail herein. Additionally, some well-known structures or functions may not be shown or described in detail below, so as to avoid unnecessarily obscuring the relevant description.

The terminology used below is to be interpreted in its broadest reasonable manner, even though it is being used in conjunction with a detailed description of certain specific examples of the facility. Indeed, certain terms may even be emphasized below; however, any terminology intended to be interpreted in any restricted manner will be overtly and specifically defined as such in this Detailed Description section.

In some cases, the system uses textual gestalt information to uniquely identify location or ranges in text, whether layout information is known or not. Some of the information includes:

• • Continuum from arrangement of articles, headlines, captions, images, etc. on a page or spread, to OCR-ed text or image match of small region of text.
    • Text-based
      • Use of character and/or word sequence
      • Use of vertical alignment of characters and/or words
    • Character Density Based
      • Use of density of character sequence
      • Use of vertical or orthogonal character densities
      • Use of paired densities
    • Image-based
      • Word lengths and/or word spaces
        • Use of relative or absolute word lengths in sequence
        • Use of relative or absolute word lengths and vertical alignment
        • Use of relative or absolute word lengths and vertical alignment seeing the vertical dimension as a sequence (disregarding line spacing details)
        • Use of relative or absolute word lengths and their sequencing, line-to-line
        • Use of relative or absolute word lengths and their proximity in sequence, especially when line breaks are not accurately known
      • Textual Features
        • Use of sequences of textual features a the word and/or character level—especially easily and reliably identifiable features such as vertical components, straight components, curved components, ascenders, descenders, density, moments (e.g. word centroids, sequences of vertical centroids), frequency (1St derivative), relative frequency (2nd derivative), edge-related features (e.g. edge density, connected-edge lengths, connected-edge region sizes, connected-edge region shapes), stroke-related features (stroke width, stroke width variance), serif-related features, line-related features (e.g. angles, curvature, line-joins (e.g. miter, rounded, beveled), line-caps (e.g. butt, round, projected square).
        • Use of textual features at the line, paragraph, article, column, region, page, spread-level.
  • Special Techniques
    • Use of blurred version of original text and generating blurred version of captured image(s) to match the blur-level of the indexed copy.
      • This optimizes the set of features extracted and indexed to match what will be encountered by a wide range of capture hardware, including cameras with no close-up focus capabilities.
    • Iterative Processing and Context
      • Use of context and inputs to generate a first estimated output at some level (e.g. estimated line height, character size, word space size, column width, etc.), then using this estimate to derive a second output, such as the word-space centroid-and-width pattern of a paragraph, then revising the first output, such as the estimated word space size, based on the likehood of the second estimate in the expected corpus of possibilities.
      • Use of such iteration in navigating a tree of possibilities, from revising the context and corpus expectations (e.g. maybe the input is from a different customer, in a different country, in a different script, etc.) to realizing that a part of an image is an “e” even though it looks more like a “c”.
      • Typically, input is expected to come from the same context as previous input, e.g. on the same page, in the same newspaper, etc. This provides strong bottom-up constraints, allowing few indications to quickly narrow in on an exact location or range. However, the iterative processing allows any given context to be falsified/rejected, for exploration of a larger context, etc. For example, a context may dictate that a character be interpreted in a certain way, but if too much counter-evidence is encountered, the context may be questioned. Similarly, multiple contexts may be explored, based on available time, processing, expense, etc., coupled with the degree to which a particular context seems questionable.

FIGS. 1-6 are document diagrams showing many different types of information derived from the spaces between words, including distance between two spaces, a triangle of distances between three spaces, many distances between a space and surrounding spaces, and so on.

Thus, the system may, in some cases, perform or employ an algorithm, such as a routine stored in memory of the system (local memory or at a remote server, for example), that performs some or all of the following: (1) identify gestalt information from document, (2) compare the gestalt information to an indexed associated with the document or a group of documents, and (3) identify the document and the location in the document based on the matching. In some cases, the system may use probabilistic methods to perform the comparison

The following discussion provides a brief, general description of a representative environment in which the facility can be implemented. Although not required, aspects of the facility may be described below in the general context of computer-executable instructions, such as routines executed by a general-purpose data processing device (e.g., a server computer, a personal computer, or mobile/portable device). Those skilled in the relevant art will appreciate that the facility can be practiced with other communications, data processing, or computer system configurations, including: wireless devices, Internet appliances, hand-held devices (including personal digital assistants (PDAs)), wearable computers, all manner of cellular or mobile phones, multi-processor systems, microprocessor-based or programmable consumer electronics, set-top boxes, network PCs, mini-computers, mainframe computers, and the like. Indeed, the terms “computer,” “server,” and the like are used interchangeably herein, and may refer to any of the above devices and systems.

While aspects of the facility, such as certain functions, are described as being performed exclusively on a single device, the facility can also be practiced in distributed environments where functions or modules are shared among disparate processing devices. The disparate processing devices are linked through a communications network, such as a Local Area Network (LAN), Wide Area Network (WAN), or the Internet. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Aspects of the facility may be stored or distributed on tangible computer-readable media, including magnetically or optically readable computer discs, hard-wired or preprogrammed chips (e.g., EEPROM semiconductor chips), nanotechnology memory, biological memory, or other data storage media. Alternatively, computer implemented instructions, data structures, screen displays, and other data related to the facility may be distributed over the Internet or over other networks (including wireless networks), on a propagated signal on a propagation medium (e.g., an electromagnetic wave(s), a sound wave, etc.) over a period of time. In some implementations, the data may be provided on any analog or digital network (packet switched, circuit switched, or other scheme).

The term “phone,” as used herein, may be a cell phone, a personal digital assistant (PDA), a portable email device (e.g., a Blackberry® device), a portable media player (e.g., an Apple IPod Touch®), or any other device having wireless communication capability. The network may be any type of cellular, IP-based, converged telecommunications network, etc., including but not limited to Global System for Mobile Communications (GSM), Time Division Multiple Access (TDMA), Code Division Multiple Access (CDMA), Orthogonal Frequency Division Multiple Access (OFDMA), General Packet Radio Service (GPRS), Enhanced Data GSM Environment (EDGE), Advanced Mobile Phone System (AMPS), Worldwide Interoperability for Microwave Access (WiMAX), Universal Mobile Telecommunications System (UMTS), Evolution-Data Optimized (EVDO), Long Term Evolution (LTE), Ultra Mobile Broadband (UMB), Voice over Internet Protocol (VoIP), Unlicensed Mobile Access (UMA), wireless fidelity (WiFi), any of the IEEE 802.11x protocols, etc.

Many, if not all, displays of information (e.g., text printed or rendered on a document, images displayed by a laptop screen, advertisements on a billboard) are associated with electronic content. The electronic content may be an electronic counterpart to a paper document (e.g., the online version of a printed newspaper article), or associated electronic content (e.g., a movie based on a book or a Wikipedia entry for a phrase). Often, there exists a discrete amount of information in the display that can identify the electronic counterpart or associated content, and capturing a sample of the information (such as a text fragment) using a capture device is enough to identify and locate the electronic counterpart and/or associated content. In most cases, the amount of text needed is very small. That is, a few words of text from a document can function as an identifier for the document and link to the electronic counterpart and/or associated content. Also, the few words may identify not only the document, but also a location within the document. The system described herein associates these displays of information and their digital counterparts and associated content to enable users to capture and interact with the information in many useful ways.

Simply, when a user captures, images, or scans a few words, characters, or regions in a paper document, the system can retrieve the electronic counterpart document or some part of it, display the electronic counterpart or some part of it, email it to somebody, purchase it, print it, post it to a web page, or perform other actions that enable a user to interact with the document or related content. For example, a user hovers his/her mobile device (and its camera) over a portion of a newspaper article, causing the user's mobile device to display an electronic version of the article on the touch screen of the mobile device, as well as provide a menu of options to the user that allow the user to further interact with the article.

The system implements these and many other examples of “paper/digital integration” without requiring changes to the current processes of writing, printing and publishing documents, giving such conventional rendered documents a whole new layer of digital functionality. FIG. 7 is a data flow diagram that illustrates the flow of information in some examples of a suitable system. Other examples may not use all of the stages or elements illustrated here, while some will use more.

FIG. 7 and the following discussion provide a brief, general description of a suitable environment in which the system can be implemented. Although not required, aspects of the system are described as computer-executable instructions, such as routines executed by a general-purpose data processing device, e.g., a server computer, wireless device, mobile phone or personal computer. Those skilled in the relevant art will appreciate that aspects of the system can be practiced with other communications, data processing, or computer system configurations, including: Internet appliances, hand-held devices (including personal digital assistants (PDAs)), all manner of cellular or mobile phones, telecommunications or mobile devices, multi-processor systems, microprocessor-based or programmable consumer electronics, set-top boxes, network PCs, mini-computers, mainframe computers, and the like. Indeed, the terms “computer,” “server,” and the like are generally used interchangeably herein, and refer to any of the above devices and systems, as well as any data processor.

Aspects of the system can be embodied in a special purpose computer or data processor that includes components specifically programmed, configured, or constructed to perform one or more of the computer-executable instructions explained in detail herein. While aspects of the system, such as certain functions, are described as being performed exclusively on a single device, the system can also be practiced in distributed environments where functions or modules are shared among disparate processing devices, which are linked through a communications network, such as a Local Area Network (LAN), Wide Area Network (WAN), or the Internet. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.

Aspects of the system may be stored or distributed on tangible computer-readable storage media, including magnetically or optically readable computer discs, hard-wired or preprogrammed chips (e.g., EEPROM semiconductor chips), secure memory (such as SIM cards) nanotechnology memory, biological memory, or other data storage media. Alternatively, computer implemented instructions, data structures, screen displays, and other data under aspects of the system may be distributed over the Internet or over other networks (including wireless networks), or they may be provided on any analog or digital network (packet switched, circuit switched, or other scheme).

Referring to FIG. 7, a capture device, such as a mobile device having a camera and/or voice recorder, captures 700 text and/or other information from a rendered document, and processes 702 the captured data, for example to remove artifacts of the capture process, to improve the signal-to-noise ratio, and so on. The system, via a recognition component (such as OCR device, speech recognition device, or autocorrelation device or a component that employs the system described herein) then optionally converts 704 the data into one or more signatures, such as segments of text, text offsets, or other symbols or characters, or the system described herein Alternatively, the system performs an alternate form of extracting one or more document signatures from the rendered document. In some cases, the signature represents a set of possible text transcriptions. In some cases, the process may be influenced or constrained by feedback from other previously or subsequently performed steps. For example, the system may previously identify candidate documents from which the capture likely originates, thus narrowing the possible interpretations of the original capture.

Post processing components may receive data from the recognition process and filter 706 the data, or perform other operations, as desired. In some examples, the system may deduce, determine, identify, and/or perform direct actions 707 immediately and without proceeding to the following steps in the routine, such as when the system captures a phrase or symbol that contains sufficient information to convey the user's intent. In these cases, the system may not need to identify or reference a digital counterpart document in order to carry out the user's wishes.

The system, in step 708, may then construct a query or a set of queries for use in searching for an electronic counterpart or other content associated with the capture. Some aspects of the query construction may depend on the search process used, and the system may perform them in a later step (such as after a search is performed), but there will typically be some operations, such as the removal of obviously misrecognized or irrelevant characters, the system can perform in advance.

The system passes 710 the query or queries to a search and context analysis component. The system may attempt to identify the document from which the original data was captured. To do so, the system may use search indices and search engines 712, knowledge about the user 714 and/or knowledge about the user's context or the context in which the capture occurred 716. For example, the system may interact with a search engine 712 that employs and/or indexes information specifically about rendered documents, about their digital counterpart documents, and/or about documents that have a web (internet) presence). The system may transfer information back and forth with these information sources, and may feed identified information into various other steps of the routine. For example, the system may receive information about the language, font, rendering and likely next words of a capture based on receiving knowledge of candidate documents during step 710.

The system, in step 720, may retrieve a copy of the document or documents identified earlier as being electronic counterparts to the rendered document. The system may have direct access to document sources and repositories 724 (e.g., a local filing system or database or a web server), or the system may contact an access service 722 to retrieve a document or documents. The access service 722 may enforce authentication, security or payments for documents, or may provide other services, such as conversion of the document into a desired format, among other things.

Applications of the system may take advantage of the association of extra functionality or data with part or all of a document. For example, advertising applications may associate particular advertising messages or subjects with portions of a document, such as keywords, phrases, or proximities to certain content. This extra associated functionality or data may be thought of as one or more overlays on the document, and is referred to herein as “markup.” Thus, in step 730, the system identifies any markup relevant to the captured data and/or an identified electronic counterpart. In some cases, the markup is provided by the user, the originator, the publisher of the document, other users of the document, and so on, and may be stored at a directly accessible source 732, or dynamically generated by a markup service 734. In some examples, the markup can be associated with, and apply to, a rendered document and/or the digital counterpart to a rendered document, or to groups of either or both of these documents.

As a result of some or all of the previous steps, the system may take or perform 740 actions. The actions may be system default actions, such as simply recording the information found, may be dependent on the data or document, or may be derived from the markup analysis. In some cases, the system may simply pass data to another system. In some cases, the possible actions appropriate to a capture at a specific point in a rendered document will be presented to the user as a menu on an associated display, such as a capture device's display (the touch screen of a mobile device) or an associated display (the user's laptop screen). The system may identify or perform an action or actions in response to the capture, in response to a user request to perform an action or actions, or a later time.

As discussed herein, a suitable system or operating environment includes a number of different components. For example, the system may include one or more optical scanning capture devices or voice capture devices (such as mobile phones and other mobile computing devices, hand-held scanning devices, and so on). The capture devices communicate with other components of the system, such as a computer or other mobile devices, using either wired or wireless connections or over a network.

The capture devices, computers and other components on the network may include memory containing computer executable instructions for processing received data or information captured from rendered documents and other sources (such as information displayed on a screen or monitor). The network may also include search engines, document sources, user account services and storage devices, markup services, and other network services. Examples of a suitable network may include a corporate Intranet, the public Internet, a mobile phone network or some other network, or any interconnection of the above.

Regardless of the manner by which the devices and components are coupled to each other, they may all may be operable in accordance with well-known commercial transaction and communication protocols (e.g., Internet Protocol (IP)). In some examples, many of the functions and capabilities of the system may be incorporated or integrated into the capture device.

As described above, the capture device may capture text using an optical scanner that captures image data from the rendered document, or using an audio recording device that captures a user's spoken reading of the text, or other methods. In some examples, the capture device may also capture images, graphical symbols and icons, and so on, including machine-readable codes such as barcodes, although these are not generally required to recognize a document or perform actions. The device may be exceedingly simple, and include little more than a transducer, some storage, and a data interface, relying on other functionality residing elsewhere in the system, or it may be a more full-featured device. In some cases, the device may be a mobile device running one or more applications that perform some or all of the functionality described herein.

The capture device includes a capture element that captures text, symbols, graphics, and so on, from rendered documents and other displays of information. The capture element may include an optical scanning head, a camera, optical sensors, and so on.

The capture device may include logic to interact with the various other components of the device. The logic may be operable to read and write data and program instructions stored in associated storage, such as RAM, ROM, flash, or other suitable memory. The capture device may also include a touch screen or other user interface to communicate information to users, and receive input from users, among other things. The capture device may include an on board power source, or may be powered over a USB or other similar connections.

As an example of how the capture device may be used, a reader may capture text from a newspaper article with a camera associated with her mobile device. The text is captured as a bit-mapped image via the camera. The logic stores the bit-mapped image in memory and time stamps the image. The logic also performs optical character recognition (OCR), and converts the image to text. The system uploads the text to an index of content associated with the newspaper, and identifies and retrieves an electronic counterpart for the article. The capture device then displays the electronic counterpart via an associated touch screen along with one or more actions to perform.

Further details regarding system processes, components, and/or devices may be found in commonly assigned PCT Application No. PCT/EP/2007/008075, filed on Sep. 17, 2007, entitled CAPTURE AND DISPLAY OF ANNOTATIONS IN PAPER AND ELECTRONIC DOCUMENTS, and U.S. Provisional Patent Application No. 61/153,614, filed on Feb. 18, 2009, entitled DOCUMENT INTERACTION SYSTEM AND METHOD, which are hereby incorporated by reference in their entirety.

Conclusion

Unless the context clearly requires otherwise, throughout the description and the claims, the words “comprise,” “comprising,” and the like are to be construed in an inclusive sense, as opposed to an exclusive or exhaustive sense; that is to say, in the sense of “including, but not limited to.” As used herein, the terms “connected,” “coupled,” or any variant thereof, means any connection or coupling, either direct or indirect, between two or more elements; the coupling of connection between the elements can be physical, logical, or a combination thereof. Additionally, the words “herein,” “above,” “below,” and words of similar import, when used in this application, shall refer to this application as a whole and not to any particular portions of this application. Where the context permits, words in the above Detailed Description using the singular or plural number may also include the plural or singular number respectively. The word “or,” in reference to a list of two or more items, covers all of the following interpretations of the word: any of the items in the list, all of the items in the list, and any combination of the items in the list.

The above detailed description of examples of the system is not intended to be exhaustive or to limit the system to the precise form disclosed above. While specific examples of, and examples for, the system are described above for illustrative purposes, various equivalent modifications are possible within the scope of the system, as those skilled in the relevant art will recognize. For example, while processes or blocks are presented in a given order, alternative examples may perform routines having steps, or employ systems having blocks, in a different order, and some processes or blocks may be deleted, moved, added, subdivided, combined, and/or modified to provide alternative or subcombinations. Each of these processes or blocks may be implemented in a variety of different ways. Also, while processes or blocks are at times shown as being performed in series, these processes or blocks may instead be performed in parallel, or may be performed at different times. Further any specific numbers noted herein are only examples: alternative implementations may employ differing values or ranges. Those skilled in the art will also appreciate that the actual implementation of a database may take a variety of forms, and the term “database” is used herein in the generic sense to refer to any data structure that allows data to be stored and accessed, such as tables, linked lists, arrays, etc.

While certain aspects of the system are presented below in certain claim forms, the inventors contemplate the various aspects of the system in any number of claim forms. For example, while only one aspect of the system is recited as embodied in a computer-readable medium, other aspects may likewise be embodied in a computer-readable medium. Accordingly, the inventors reserve the right to add additional claims after filing the application to pursue such additional claim forms for other aspects of the system.

From the foregoing, it will be appreciated that specific embodiments of the facility have been described herein for purposes of illustration, but that various modifications may be made without deviating from the spirit and scope of the facility. Accordingly, the facility is not limited except as by the appended claims.

Claims

1. A computer-implemented method comprising: obtaining an image of a rendered document that includes text;determining a two dimensional geometric shape based at least on a first location in the image of the rendered document of a first space between a first pair of consecutive words, a second location in the image of the rendered document of a second space between a second pair of consecutive words, and a third location in the image of the rendered document of a third space between a third pair of consecutive words, wherein the first space, the second space, and the third space are not all included on a same line of text in the rendered document;generating a document signature based on the two dimensional geometric shape; andgenerating a query for an electronic document that is a counterpart to the rendered document, based at least on the document signature.

2. The method of claim 1, wherein the first location, the second location, and the third location each comprise an absolute location in the image and determining the two dimensional geometric shape comprises determining the two dimensional geometric shape using the absolute locations.

3. The method of claim 1, comprising: determining a first distance between the first space and the second space;determining a second distance between the first space and the third space; anddetermining a third distance between the second space and the third space,wherein determining the two dimensional geometric shape comprises determining the two dimensional geometric shape using the first distance, the second distance, and the third distance.

4. The method of claim 3, wherein the two dimensional geometric shape comprises a triangle and the document signature indicates a signature location of the triangle in the image.

5. The method of claim 3, wherein the two dimensional geometric shape comprises a word centroid and the document signature indicates a signature location of the centroid in the image.

6. The method of claim 5, wherein the signature location comprises a location of the center of the centroid in the image.

7. The method of claim 1, wherein the two dimensional geometric shape is defined by a first line that extends from the first location to the second location, a second line that extends from the second location to the third location, and a third line that extends from the third location to the first location.

8. The method of claim 1, wherein the two dimensional geometric shape comprises an outline that connects the first location, the second location, and the third location.

9. The method of claim 1, further comprising comparing the two dimensional shape to two dimensional shapes of other documents to identify a matching document that matches the rendered document.

10. A system comprising: one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:obtaining an image of a rendered document that includes text;determining a two dimensional geometric shape based at least on a first location in the image of the rendered document of a first space between a first pair of consecutive words, a second location in the image of the rendered document of a second space between a second pair of consecutive words, and a third location in the image of the rendered document of a third space between a third pair of consecutive words, wherein the first space, the second space, and the third space are not all included on a same line of text in the rendered document;generating a document signature based on the two dimensional geometric shape; andgenerating a query for an electronic document that is a counterpart to the rendered document, based at least on the document signature.

11. The system of claim 10, wherein the first location, the second location, and the third location each comprise an absolute location in the image and determining the two dimensional geometric shape comprises determining the two dimensional geometric shape using the absolute locations.

12. The system of claim 10, the operations comprising: determining a first distance between the first space and the second space;determining a second distance between the first space and the third space; anddetermining a third distance between the second space and the third space,wherein determining the two dimensional geometric shape comprises determining the two dimensional geometric shape using the first distance, the second distance, and the third distance.

13. The system of claim 12, wherein the two dimensional geometric shape comprises a triangle and the document signature indicates a signature location of the triangle in the image.

14. The system of claim 12, wherein the two dimensional geometric shape comprises a word centroid and the document signature indicates a signature location of the centroid in the image.

15. The system of claim 14, wherein the signature location comprises a location of the center of the centroid in the image.

16. A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising: obtaining an image of a rendered document that includes text;determining a two dimensional geometric shape based at least on a first location in the image of the rendered document of a first space between a first pair of consecutive words, a second location in the image of the rendered document of a second space between a second pair of consecutive words, and a third location in the image of the rendered document of a third space between a third pair of consecutive words, wherein the first space, the second space, and the third space are not all included on a same line of text in the rendered document;generating a document signature based on the two dimensional geometric shape; andgenerating a query for an electronic counterpart to the rendered document, based at least on the document signature.

17. The computer-readable medium of claim 16, wherein the first location, the second location, and the third location each comprise absolute location in the image and determining the two dimensional geometric shape comprises determining the two dimensional geometric shape using the absolute locations.

18. The computer-readable medium of claim 16, the operations comprising: determining a first distance between the first space and the second spacedetermining a second distance between the first space and the third space; anddetermining a third distance between the second space and the third space,wherein determining the two dimensional geometric shape comprises determining the two dimensional geometric shape using the first distance, the second distance, and the third distance.

19. The computer-readable medium of claim 18, wherein the two dimensional geometric shape comprises a triangle and the document signature indicates a signature location of the triangle in the image.

20. The computer-readable medium of claim 18, wherein the two dimensional geometric shape comprises a word centroid and the document signature indicates a signature location of the centroid in the image.

21. The computer-readable medium of claim 20, wherein the signature location comprises a location of the center of the centroid in the image.