Patent Application
20080123895
1. Field of the Invention
The technology disclosed herein relates to methods of operating explosive detection systems and medical imaging systems generally, and more particularly, to a method and system for fast volume cropping of air volumetric data from image data of an object.
2. Discussion of Related Art
Various government agencies throughout the world are responsible for quickly and accurately identifying contraband and/or dangerous materials within passenger baggage in ways that minimize passenger inconvenience and travel times. Countless x-ray baggage scanning systems used by such agencies are of the “line scanner” type. This type of x-ray baggage scanning system includes a conveyor belt, a stationary x-ray source, and a stationary linear detector array. The conveyor belt transports a bag (e.g., a single piece of passenger baggage), through the scanner and between the stationary x-ray source and the stationary detector array. The x-ray source produces an x-ray beam. As the bag moves into the x-ray beam, the beam passes through and is partially attenuated by the bag, before being received by the detector array. Each two-dimensional region of the baggage through which the x-rays penetrate forms a planar segment (“slice”) having a unique density of x-rays that varies depending on how much attenuation the baggage affords. Each time the x-ray source activates, each detector of the detector array records projection data, which is conventionally calculated as the integral of the density of each planar segment of the baggage. Once obtained, the projection data is processed by a computer and used to reconstruct a two-dimensional density image of the baggage. Customarily, the two-dimensional density image of the baggage is displayed for analysis by a human operator.
Other types of x-ray baggage scanning systems use various types of x-ray computed tomography (CT) to identify objects within baggage that is conveyed through the scanning system. Manufacturers of x-ray CT scanning systems include GE Homeland Protection, Inc. (formerly InVision, Inc.) of Newark, Calif., which is a subsidiary of the General Electric Company, and International Security Systems Corporation, which is a subsidiary of Analogic Corporation.
One example of a conventional CT scanner is a dual-energy, helical cone beam, multi-slice CT scanner, developed by Analogic, Inc., which can generate 3-D image data of all objects in a bag, collect all image data in one pass, automatically analyze the entire contents of the bag, and scan up to six hundred bags per hour. In such a CT system, sophisticated software can automatically isolate, analyze, and evaluate bag contents against the known characteristics of explosives, illegal drugs, and other contraband. If a match is found, the CT scanning system operator is notified, the area of concern is highlighted, and/or a full rotating three-dimensional image of the potential threat is provided for further analysis.
Additionally, explosive detection systems that integrate multiple types of scanning systems have been developed. One example is an advanced technology explosive detection system developed by GE Homeland Protection, Inc. (formerly InVision, Inc.) of Newark, Calif., which is a subsidiary of the General Electric Company that combines a coherent x-ray scatter (CXRS) scanner with a CT scanner and offers data-fusion between the scanners for baggage screening. When used in the scanner-fused explosive detection system, the CXRS scanner uses alarm location data, acquired by the CT scanner positioned earlier in the baggage handling system, to limit the CXRS scan to specific areas of bags that the CT scanner previously identified as suspicious. CXRS uses molecular composition to identify alarm objects, and is used in combination with x-ray CT to lower false alarm rates and improve baggage throughput.
Although x-ray CT scanners are useful, their current methods of operation share a common disadvantage, which is that a significant portion of the image volumetric data obtained by the x-ray CT scanner consists of air volumetric data that borders object volumetric data (e.g., a bag, clothing, shoe, etc.). This is natural since the cross-sectional diameter of the object of interest is typically smaller than the diameter of the rotatable gantry that forms part of a conventional x-ray CT scanner. Various methods have been proposed to solve this problem, but are undesirable due to the immense amounts of raw computer processing power and scan times required. One such method utilizes per-voxel iteration through all of the image data. Another method moves slices in from the edges of the image data and searches them for voxel values that exceed a pre-determined threshold. Another method uses a multi-dimensional bisection approach. Yet another method uses a ray-casting approach. A drawback of such methods is that digital image processing in three dimensions (3D) using such methods is computationally expensive. Moreover, processing the significant amounts of air volumetric data associated with such methods is wasteful and time-consuming for high-volume imaging systems.
A solution is thus needed that provides a method, applicable to x-ray CT scanners, and other imaging systems, that minimizes the number of voxels to be analyzed in the image data of an object (e.g., a bag, a medical patient, a product, etc.). It is further desired that such a solution be easily implemented in existing imaging systems used in security, medical, engineering, and other types of applications. Such a solution can yield reduced processing times for passenger baggage, medical patients, product inspection/testing, etc., offering the potential for reduced operating costs.
Embodiments of the invention overcome the disadvantages associated with the related art and meet the needs discussed above by providing a novel method and system for identifying, cropping, and (optionally) discarding air volumetric data from around volumetric object data in image data obtained by an x-ray CT scanner or other imaging device. Such a method is relatively simple, cost-effective, and efficient. It also significantly lessens the amount of computer processing required and reduces scan times by focusing the subsequent imaging and/or threat detection analysis on only the volumetric object data that remains after the air volumetric data has been quickly identified and cropped. Embodiments of the novel method are suitable for use in security applications, medical applications, engineering applications, etc. It also is more efficient for network transmissions and disk storage.
Technical effects afforded by embodiments of the invention include, but are not limited to, air volumetric data cropped from image data of an object; a substantial reduction in the time it takes to process the object volumetric data that remains after the air volumetric data is cropped; and a substantial increase in disk space savings, as compared to prior image processing methods and systems. The significant improvements in processing time and disk space savings result, in part, from quickly locating the boundaries of the imaged object, and from storing and processing the object volumetric data together with minimal or no air volumetric data.
In some embodiments, a method is provided. The method may comprise obtaining image data of an object from an imaging system, wherein the image data comprises air volumetric data and object volumetric data. The method may further include a step of sampling the image data in three dimensions. The method may further include a step of identifying one or more candidate voxels; and a step of identifying, from the one or more candidate voxels, one or more starting voxels.
As an alternative to the above-described embodiments, a system is provided. The system may be an explosive detection system, a medical imaging system, or an engineering imaging system. An embodiment of the system may comprise an imaging system configured to obtain image data of an object. The image data comprises air volumetric data and object volumetric data. The system may further comprise a computer processor coupled with the imaging system, and a memory readable by the computer processor. Computer executable instructions stored in the memory may also form part of the explosive detection system. When executed by the computer processor, the computer executable instructions cause the computer processor to: operate the imaging system to obtain the image data of the object; sample the image data in three dimensions; identify one or more candidate voxels; and identify, from the one or more candidate voxels, one or more starting voxels.
The foregoing has outlined rather broadly the features of the invention so that the following detailed description may be better understood. Additional features and advantages of various embodiments of the invention that form the subject matter of the appended claims may be described hereinafter.
For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:
Reference is made herein to the accompanying drawings briefly described above, which show by way of illustration various embodiments of the invention. Persons of ordinary skill in the above-referenced technological field will recognize that other embodiments may be utilized, and that various changes may be made without departing from the scope of the claimed invention.
As used herein, an element or step recited in the singular and proceeded with the word “a” or “an” includes plural elements or steps, unless exclusion of such plural elements or steps is explicitly recited.
The following terms (alarm object, bag, baggage, imaging system, object, image data, reconstructing an image, image reconstruction, object volumetric data, air volumetric data, crop, bounding box, and sampling) used herein should generally be construed as having the following exemplary meanings. As used herein the term:
“alarm object” refers to any substance or thing that has been designated for detection by an inspection system (non-limiting examples include explosives, illegal drugs, hazardous substances, tissue anomalies, product components, and the like);
“bag” refers to a piece of baggage;
“baggage” refers to all of a passenger's or traveler's personal belongings, whether checked or unchecked;
“imaging system” refers to a device that analyzes an object and converts analog information about the object into digital data.
“object” refers to anything that can be imaged by a scanner (Non-limiting examples of an “object” include all or portions of a bag, a medical patient, a commercial product, etc.);
“image data” refers to all the converted digital data (whether in viewable or non-viewable form) that corresponds to the analog information about the scanned object, and includes both object volumetric data and air volumetric data;
“object volumetric data” refers to three-dimensional digital data that corresponds to the object itself,
“air volumetric data” refers to three-dimensional digital data that corresponds to the air surrounding the object;
“crop” refers to removing unwanted parts (e.g., air volumetric data) from image data;
“bounding box” refers to a closed volume that contains all, or substantially all (e.g., greater than about 95% of), the object volumetric data (In some embodiments, inner surfaces of the closed volume may define or approximate the three-dimensional outer surfaces of the object);
“sampling” refers to selecting a subset of image data voxels for comparison to a predetermined voxel threshold; and
“voxel threshold” refers to a scan setting that determines a point at which a voxel will be categorized as an “air voxel” or an “object voxel” depending upon its scanned color and/or density value.
Although medical and engineering embodiments are not depicted in the Figures, the imaging system 102 may be an x ray CT scanner configured for use in medical or engineering applications. In medical applications, the object that is scanned may be a medical patient (or a portion thereof), and the alarm object may be a tissue irregularity such as a blood clot, a tumor, and the like. In engineering applications, the object may be a manufactured product (or portion thereof) that is being tested and/or reverse engineered. In an embodiment directed to medical applications, the imaging system 102 may be a magnetic resonance imaging (MRI) scanner, or other type of medical imaging system.
In the embodiment shown in
Referring to
The method 300 may further include a step 302 of sampling the image data in at least one dimension (and preferably two or more dimensions) until at least one voxel (and preferably more voxels) fall above a predetermined minimum voxel threshold that has been selected to distinguish object voxels from air voxels. In an optional embodiment, the sampling of the image data in three dimensions may occur randomly. A voxel that falls above the predetermined minimum voxel threshold may hereinafter be referred to as “a candidate voxel.” The maximum and minimum (in X, Y, and Z, respectively) candidate voxels may hereinafter be referred to as “starting voxels” since they are used, in one embodiment of the invention, as starting points for a voxel-by-voxel iteration of a slice in at least one of the positive and negative X, Y, and Z dimensions. It will be appreciated that this predetermined minimum voxel threshold will vary depending upon the type of object imaged, the type of imaging apparatus used, and the like. In an embodiment, a step of identifying one or more candidate voxels may include comparing values of voxels that comprise the sampled image data with the predetermined voxel threshold.
The method 300 may further include a step 303 of identifying, for the one or more starting voxels from step 302 that are above the predetermined voxel threshold, a minimum and a maximum X, Y, and Z coordinate. In other words, the method 300 may identify each of the one or more starting voxels based on at least one of a maximum and minimum coordinate in three dimensions. Embodiments of steps 301, 302, 303 are further described below with respect to
Referring again to
The method 300 may further include a step 308 of creating, for a minimum X voxel, a YZ slice that passes through the minimum X voxel. The method 300 may further include a step 309 of decreasing the X position of the YZ slice by at least one voxel-width (in the negative X direction). The method 300 may further include a step 310 of determining whether the number of voxels in the iterated YZ slice that are above the predetermined voxel threshold exceed a predetermined amount. If yes, the method 300 loops back to step 309, and the YZ slice is iterated, voxel-by-voxel, until the number of voxels in the iterated YZ slice that are above the predetermined voxel threshold do not exceed the predetermined amount. At step 311, an X value of this last YZ slice is stored in a computer readable medium as a “Minimum X” value for the bounding box.
Embodiments of steps 304 to 311 are further described below with respect to
Referring to
The method 300 may further include a step 316 of creating, for a minimum Y voxel, a XZ slice that passes through the minimum Y voxel. The method 300 may further include a step 317 of decreasing the Y position of the XZ slice by at least one voxel width (in the negative Y direction). The method 300 may further include a step 318 of determining whether the number of voxels in the iterated XZ slice that are above the predetermined voxel threshold exceed a predetermined amount. If yes, the method 300 loops back to step 316, and the XZ slice is iterated, voxel-by-voxel, until the number of voxels in the iterated XZ slice that are above the predetermined voxel threshold do not exceed the predetermined amount. At step 319, a Y value of this last XZ slice is stored in a computer readable medium as a “Minimum Y” value for the bounding box.
Embodiments of steps 312 to 319 are further described below with respect to
Referring to
The method 300 may further include a step 324 of creating, for a minimum Z voxel, a XY slice that passes through the minimum Z voxel. The method 300 may further include a step 325 of decreasing the Z position of the XY slice by at least one voxel width (in the negative Z direction). The method 300 may further include a step 326 of determining whether the number of voxels in the iterated XY slice that are above the predetermined voxel threshold exceed a predetermined amount. If yes, the method 300 loops back to step 325, and the XY slice is iterated, voxel-by-voxel, until the number of voxels in the iterated XY slice that are above the predetermined voxel threshold do not exceed the predetermined amount. At step 327, a Z value of this last XY slice is stored in a computer readable medium as a “Minimum Z” value for the bounding box.
Referring to
The method 300 may further include a step 330 of storing the object volumetric data (and/or a small amount of air volumetric data, if any, that falls within the limits of the bounding box) that remains after step 329 in a computer readable memory. Optionally, the method 300 may include a step 331 of displaying the attained object volumetric data on a display device. In an alternative embodiment, the method 300 may optionally include a step of processing the object volumetric data within the bounding box for a presence of an alarm object. In another embodiment, the method 300 may optionally include a step of providing the object volumetric data within the bounding box for further processing, such as, but not limited to, explosives detection.
In an embodiment, one or more steps of the method 300 are implemented in a computer processor and associated memory elements within a medical, engineering, or security imaging system, for example, within the integrated explosive detection system 103 of
Referring to
Referring to
Referring to
Referring to
Referring to
Referring to
As mentioned above with respect to steps 320-327 of the method 300, the processes of forming slices, iterating slices, and storing minimum and maximum axis values are repeated in Z (not shown).
A detailed description of various embodiments of the claimed invention has been provided; however, modifications within the scope of the claimed invention will be apparent to persons having ordinary skill in the above-referenced technological field. Such persons will appreciate that features described with respect to one embodiment may be applied to other embodiments. Thus, the scope of the claimed invention is to be properly construed with reference to the following claims.
