Collation device, collation method, and recording medium

Abstract

A collation device according to the present invention comprises: a first calculation means for calculating a first degree of similarity by using respective pieces of target partial feature information of at least two target partial regions in a target biological image and respective pieces of registered partial feature information of at least two registered partial regions in a registered biological image; a second calculation means for calculating a second degree of similarity by using entire target feature information of the entire target biological image and entire registered feature information of the entire registered biological image when the first degree of similarity is greater than or equal to a first threshold; and a determination means for outputting a collation result, of the target biological image and the registered biological image, based on at least the second degree of similarity.

Description

Verification device, verification method, and recording medium

[0001] The present disclosure relates to the technical fields of a verification device, a verification method, and a recording medium.

[0002] As a technology for matching biometric images, for example, a technology has been proposed in which a first similarity is calculated using a first feature point extracted from a target biometric image and a second feature point of multiple registered biometric images, the registered biometric images are narrowed down based on the calculated first similarity, a second similarity is calculated using a third feature point extracted from the target biometric image and a fourth feature point of the narrowed down registered biometric images, and the registered biometric image is identified based on the calculated second similarity (see Patent Document 1).

[0003] International Publication No. 2020 / 148874

[0004] An object of this disclosure is to provide a matching device, a matching method, and a recording medium that aim to improve the technology related to the above-mentioned prior art documents.

[0005] One aspect of the matching device includes a first calculation means for calculating a first similarity using each of the target partial feature information of at least two target partial regions in a target biometric image and each of the registered partial feature information of at least two registered partial regions in a registered biometric image, a second calculation means for calculating a second similarity using the target overall feature information of the entire target biometric image and the registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold, and a determination means for outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

[0006] One aspect of the matching method is a matching method executed by a computer, which includes calculating a first similarity using each of the target portion feature information of at least two target partial regions in a target biometric image and each of the registered portion feature information of at least two registered partial regions in a registered biometric image, calculating a second similarity using the target overall feature information of the entire target biometric image and the registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold, and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

[0007] One aspect of the recording medium has recorded thereon a computer program for causing a computer to execute a matching method, which includes calculating a first similarity using each of the target partial feature information of at least two target partial regions in a target biometric image and each of the registered partial feature information of at least two registered partial regions in a registered biometric image, calculating a second similarity using the target overall feature information of the entire target biometric image and the registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold, and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

[0008] FIG. 1 is a block diagram showing an example of the configuration of a matching device according to an embodiment. FIG. 2 is a flowchart showing an example of the operation of a matching device according to an embodiment. FIG. 3 is a block diagram showing an example of the configuration of a matching device according to an embodiment. FIG. 4 is a flowchart showing an example of the operation of a matching device according to an embodiment. FIG. 5 is a conceptual diagram showing an example of the operation of a matching device according to an embodiment. FIG. 6 is a block diagram showing an example of the configuration of a matching device according to an embodiment. FIG. 7 is a flowchart showing an example of the operation of a matching device according to an embodiment. FIG. 8 is a conceptual diagram showing an example of the operation of a matching device according to an embodiment. FIG. 9 is a block diagram showing an example of the configuration of a matching device according to an embodiment.

[0009] Hereinafter, embodiments of a verification device, a verification method, and a recording medium will be described with reference to the drawings. [1: First Embodiment]

[0010] A first embodiment of a verification device, a verification method, and a recording medium will be described with reference to Figures 1 and 2. In the following, the first embodiment of the verification device, the verification method, and a recording medium will be described using a verification device 10.

[0011] 1, the verification device 10 includes a first calculation unit 11, a second calculation unit 12, and a determination unit 13. The operation of the verification device 10 will be described with reference to the flowchart of FIG.

[0012] As shown in FIG. 2 , the first calculation unit 11 calculates a first similarity using feature information of at least two partial regions in a target biometric image and feature information of at least two partial regions in a registered biometric image (step S11). A partial region is a region included in a biometric image and is a local region of the biometric image. The first calculation unit 11 may acquire each of the at least two partial regions in the target biometric image. A partial region in a target biometric image may be referred to as a "target partial region." For example, the first calculation unit 11 may acquire each of at least two target partial regions obtained by dividing the target biometric image into at least two. Furthermore, each target partial region may have an overlapping portion. In other words, each target partial region may partially include a region of the same target image. The first calculation unit 11 may extract feature information from each of the at least two target partial regions. The feature information of the target partial region may be referred to as "target partial feature information." The first calculation unit 11 may acquire each of the at least two partial regions in a registered biometric image. A partial region in a registered biometric image may be referred to as a "registered partial region." The first calculation unit 11 may extract feature information from each of at least two registered partial regions. The feature information of the registered partial region may be referred to as "registered partial feature information." The first calculation unit 11 may acquire the registered partial feature information extracted from each of at least two registered partial regions. In other words, the first calculation unit 11 may calculate the similarity between the partial regions of the biometric image. The first similarity calculated by the first calculation unit 11 may indicate the similarity between the partial regions of the biometric image.

[0013] The second calculation unit 12 determines whether the first similarity is greater than or equal to a first threshold (step S12). If the first similarity is greater than or equal to the first threshold (step S12: Yes), the second calculation unit 12 calculates a second similarity using the overall feature information of the target biometric image and the overall feature information of the registered biometric images (step S13). The second calculation unit 12 may extract feature information from the entire target biometric image. The overall feature information of the target biometric image may be referred to as "target overall feature information." The second calculation unit 12 may extract feature information from the entire registered biometric image. The overall feature information of the registered biometric image may be referred to as "registered overall feature information." The second calculation unit 12 may acquire registered overall feature information extracted from the entire registered biometric image. In other words, the second calculation unit 12 calculates the overall similarity of the biometric images. The second similarity calculated by the second calculation unit 12 may indicate the similarity between the entire biometric images.

[0014] The determination unit 13 compares the target biometric image with the registered biometric image based on at least the second similarity (step S14). If at least the second similarity indicates that the target biometric image and the registered biometric image are more similar than a predetermined standard, the determination unit 13 may determine that the target biometric image and the registered biometric image are biometric images of the same individual.

[0015] The determination unit 13 outputs the result of the matching (step S15). The determination unit 13 may output information indicating whether the target biometric image and the registered biometric image are biometric images of the same individual.

[0016] In this way, the matching device 10 performs a matching method that includes calculating a first similarity using each of the target portion feature information of at least two target partial regions in the target biometric image and each of the registered portion feature information of at least two registered partial regions in the registered biometric image, calculating a second similarity using the target overall feature information of the entire target biometric image and the registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold, and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

[0017] The above-described matching device 10 may be realized by a computer reading a computer program recorded on a recording medium. In this case, the computer program may cause the computer to execute a matching method including: calculating a first similarity using target portion feature information of at least two target partial regions in the target biometric image and registered portion feature information of at least two registered partial regions in the registered biometric image; calculating a second similarity using target overall feature information of the entire target biometric image and registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold; and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity. [Technical Effect]

[0018] The matching device 10 according to the present disclosure can accurately match a target biometric image with a registered biometric image by using both a similarity based on feature information of a partial region of the biometric image and a similarity based on feature information of the entire biometric image. [2: Second Embodiment]

[0019] A second embodiment of the collation device, the collation method, and the recording medium will be described with reference to Figures 3 to 5. The second embodiment of the collation device, the collation method, and the recording medium will be described below using the collation device 20. Note that, for the second embodiment, descriptions that overlap with the description of the first embodiment will be omitted as appropriate.

[0020] The target in this embodiment is a living body. The target in this embodiment may be a target for matching. The biometric image may be an image of a biometric pattern. The biometric pattern may be a pattern that can be acquired by capturing an image of a living body. Examples of biometric patterns include handprints, palm prints, finger vein patterns, palm vein patterns, toenail prints, palm prints, toe vein patterns, and sole vein patterns. In this embodiment, a case will be described in which the biometric image is a fingerprint image, which is an example of an image of a biometric pattern.

[0021] The matching device 20 may perform biometric authentication by matching a target biometric image with a registered biometric image. The matching device 20 according to this embodiment may perform fingerprint authentication by matching a target fingerprint image with a registered fingerprint image. [2-1: Configuration of the matching device 20]

[0022] The configuration of the verification device 20 will be described with reference to Fig. 3. Fig. 3 is a block diagram showing the configuration of the verification device 20.

[0023] 3, the verification device 20 includes a calculation device 21, a storage device 22, and a communication device 23. The verification device 20 may further include an input device 24 and an output device 25. However, the verification device 20 does not necessarily include at least one of the input device 24 and the output device 25. The calculation device 21, the storage device 22, the communication device 23, the input device 24, and the output device 25 may be connected via a data bus 26.

[0024] The arithmetic device 21 includes at least one processor (i.e., one processor or multiple processors) as hardware. The processor may include, for example, a processor conforming to a von Neumann computer architecture. The processor conforming to the von Neumann computer architecture may include at least one of a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). The processor may include, for example, a processor conforming to a non-von Neumann computer architecture. The processor conforming to the non-von Neumann computer architecture may include at least one of an FPGA (Field Programmable Gate Array) and an ASIC (Application Specific Circuit).

[0025] The arithmetic device 21 reads a computer program 221 including at least one of computer program code and computer program instructions. For example, the arithmetic device 21 may read the computer program 221 stored in the storage device 22. For example, the arithmetic device 21 may read the computer program 221 stored in a computer-readable, non-transitory storage medium using a storage medium reading device (not shown) included in the verification device 20. The computer program 221 read from the storage medium may be stored in the storage device 22. The arithmetic device 21 may acquire (i.e., download or read) the computer program 221 from a device (not shown) located outside the verification device 20 via the communication device 23 (or another communication device). The downloaded computer program 221 may be stored in the storage device 22.

[0026] The arithmetic device 21 executes the loaded computer program 221. As a result, logical functional blocks for executing the information processing to be performed by the verification device 20 are realized within the arithmetic device 21. In other words, the arithmetic device 21, together with the storage device 22 or the like in which the computer program 221 is recorded (in other words, together with the storage device 22 and the computer program 221 recorded in the storage device 22 or the like), can function as a controller or computer for realizing the logical functional blocks for executing the processing to be performed by the verification device 20. In other words, the at least one processor included in the arithmetic device 21, the memory (recording medium) included in the storage device 22 or the like, and the computer program 221 are configured so that the verification device 20 performs the information processing to be performed by the verification device 20.

[0027] A computational model that can be constructed by machine learning may be implemented in the computational device 21 by the computational device executing the computer program 221. An example of a computational model that can be constructed by machine learning is a computational model including a neural network (so-called artificial intelligence (AI)). In this case, learning of the computational model may include learning of parameters of the neural network (e.g., at least one of a weight and a bias). The computational device 21 may perform information processing using the computational model. In other words, the operation of performing information processing may include the operation of performing information processing using the computational model. Note that a computational model that has been constructed by offline machine learning using training data may be implemented in the computational device 21. Furthermore, the computational model implemented in the computational device 21 may be updated by online machine learning on the computational device 21. Alternatively, the calculation device 21 may perform information processing using a calculation model implemented in a device external to the calculation device 21 (i.e., a device provided outside the matching device 20) in addition to or instead of the calculation model implemented in the calculation device 21.

[0028] The recording medium for recording the computer program 221 executed by the arithmetic device 21 may be at least one of a CD-ROM, CD-R, CD-RW, flexible disk, MO, DVD-ROM, DVD-RAM, DVD-R, DVD+R, DVD-RW, DVD+RW, Blu-ray (registered trademark), or other optical disk, a magnetic medium such as a magnetic tape, a magneto-optical disk, a semiconductor memory such as a USB memory, or any other medium capable of storing a program. The recording medium may include a device capable of recording a computer program (for example, a general-purpose device or a dedicated device in which the computer program 221 is implemented in a state in which it can be executed in at least one of the forms of software and firmware). Furthermore, each process or function included in the computer program 221 may be realized by a logical processing block realized within the arithmetic device 21 when the arithmetic device 21 (i.e., processor) executes the computer program 221, or may be realized by hardware such as a predetermined gate array (FPGA (Field Programmable Gate Array), ASIC (Application Specific Integrated Circuit)) provided in the arithmetic device 21, or may be realized in a form that mixes logical processing blocks and partial hardware modules that realize some elements of the hardware.

[0029] The storage device 22 includes at least one memory capable of storing desired data. In other words, the storage device 22 includes at least one memory containing desired data. For example, the storage device 22 may store a computer program 221 executed by the arithmetic device 21. In this case, the storage device 22 (memory) may be used as the above-mentioned recording medium for recording the computer program 221 executed by the arithmetic device 21. The storage device 22 may temporarily store data used by the arithmetic device 21 when the arithmetic device 21 is executing the computer program 221. The storage device 22 may store data to be stored long-term by the verification device 20. The storage device 22 may include at least one of a RAM (Random Access Memory), a ROM (Read Only Memory), a hard disk device, a magneto-optical disk device, an SSD (Solid State Drive), and a disk array device. In other words, the storage device 22 may include a non-temporary recording medium.

[0030] The storage device 22 may store a partial feature vector storage unit 222 and a minutiae storage unit 223, which will be described later.

[0031] The communication device 23 may be capable of communicating with devices external to the verification device 20. The communication device 23 may perform wired communication or wireless communication.

[0032] The input device 24 is a device capable of accepting information input to the verification device 20 from outside. The input device 24 may include an operation device (e.g., a keyboard, a mouse, a touch panel, etc.) that can be operated by a user of the verification device 20. The input device 24 may include a recording medium reading device that can read information recorded on a recording medium that is detachable from the verification device 20, such as a USB (Universal Serial Bus) memory. Note that when information is input to the verification device 20 via the communication device 23 (in other words, when the verification device 20 acquires information via the communication device 23), the communication device 23 may function as an input device.

[0033] The output device 25 is a device capable of outputting information to the outside of the verification device 20. The output device 25 may output visual information such as text or images, auditory information such as sound, or tactile information such as vibration, as the information. The output device 25 may include, for example, at least one of a display, a speaker, a printer, and a vibration motor. The output device 25 may be capable of outputting information to a recording medium that is detachable from the verification device 20, such as a USB memory. Note that when the verification device 20 outputs information via the communication device 23, the communication device 23 may function as an output device.

[0034] FIG. 3 shows an example of logical functional blocks realized in the arithmetic device 21 to execute the matching method. As shown in FIG. 3, a first matching unit 211, a second matching unit 212, and a determination unit 213 are realized in the arithmetic device 21. The first matching unit 211 may have a partial region acquisition unit 2111, a feature vector extraction unit 2112, and a first calculation unit 2113. The second matching unit 212 may have a minutiae extraction unit 2121 and a second calculation unit 2122. The "first calculation unit 2113" is a component corresponding to the "first calculation unit 11" in the first embodiment described above, the "second calculation unit 2122" is a component corresponding to the "second calculation unit 12" in the first embodiment described above, and the "determination unit 213" is a component corresponding to the "determination unit 13" in the first embodiment described above. [2-2: Matching Method Executed by Matching Device 20]

[0035] The collation method executed by the collation device 20 will be described with reference to Fig. 4. Fig. 4 is a flowchart showing an example of the processing flow of the collation method executed by the collation device 20.

[0036] 4, the partial region acquisition unit 2111 acquires a fingerprint image to be matched (target fingerprint image) (step S21). The partial region acquisition unit 2111 acquires each of at least two partial regions (target partial regions) in the target fingerprint image (step S22).

[0037] The partial region acquisition unit 2111 may divide the target fingerprint image and acquire at least two target partial regions. For example, as illustrated in FIG. 5, the partial region acquisition unit 2111 may divide the target fingerprint image TI into four and acquire each of the four target partial regions. Below, an example will be described in which the partial region acquisition unit 2111 divides the target fingerprint image into four and acquires each of the four target partial regions. Note that dividing into four is just an example, and the target fingerprint image may be divided into other numbers. Furthermore, although an example will be described in which the partial region acquisition unit 2111 divides the target fingerprint image, the target partial regions may overlap each other. In other words, the target partial regions may partially include an area of ​​the same fingerprint image.

[0038] The feature vector extraction unit 2112 acquires feature information from each of the four target partial regions (step S23). The feature vector extraction unit 2112 may extract a feature vector as feature information from each of the four target partial regions. The feature vector extracted by the feature vector extraction unit 2112 is referred to as a "target partial feature vector." In other words, the feature vector extraction unit 2112 acquires four target partial feature vectors. In the example shown in FIG. 5, the feature vector extraction unit 2112 may acquire a first target partial feature vector T1, a second target partial feature vector T2, a third target partial feature vector T3, and a fourth target partial feature vector T4.

[0039] The feature vector extraction unit 2112 may acquire a target portion feature vector using a feature extraction model that can be constructed by machine learning. The feature extraction model may be a computational model that outputs a target portion feature vector when a target portion area of ​​a fingerprint image is input. The feature extraction model may include a convolutional neural network, and training the feature extraction model may include training parameters of the convolutional neural network (e.g., at least one of weights and biases).

[0040] The partial feature vector storage unit 222 may include partial feature vectors extracted from at least two partial regions (registered partial regions) in a registered fingerprint image (registered fingerprint image). For example, as illustrated in FIG. 5, the partial feature vector storage unit 222 may include partial feature vectors extracted from four registered partial regions obtained by dividing the registered fingerprint image RI into four. The following describes an example in which the partial feature vector storage unit 222 includes partial feature vectors extracted from four registered partial regions obtained by dividing the registered fingerprint image into four. The partial feature vector storage unit 222 may include partial feature vectors for each of a plurality of registered fingerprint images. A feature vector included in the partial feature vector storage unit 222 is referred to as a "registered partial feature vector." In other words, the partial feature vector storage unit 222 includes four registered partial feature vectors for one registered fingerprint image. In the example shown in Figure 5, the partial feature vector memory unit 222 may include, for one registered fingerprint image RI, a first registered partial feature vector R1, a second registered partial feature vector R2, a third registered partial feature vector R3, and a fourth registered partial feature vector R4.

[0041] The first calculation unit 2113 acquires the registered partial feature vectors from the partial feature vector storage unit 222 (step S24). For example, if there are 10,000 registered fingerprint images, the first matching unit 211 may acquire each of the registered partial feature vectors for the 10,000 registered fingerprint images.

[0042] The first calculation unit 2113 calculates the partial similarity between each of the target partial feature vectors and each of the registered partial feature vectors (step S25). In other words, the first calculation unit 2113 calculates the similarity between partial regions of the fingerprint image. For example, if there are 10,000 registered fingerprint images, the first calculation unit 2113 may calculate the partial similarity between each of the target partial feature vectors and each of the registered partial feature vectors of the 10,000 registered fingerprint images. The first calculation unit 2113 may calculate, for example, the cosine similarity between the target partial feature vector and the registered partial feature vector to calculate each of the partial similarities. Alternatively, the first calculation unit 2113 may calculate the Euclidean distance between the target partial feature vector and the registered partial feature vector to calculate each of the partial similarities.

[0043] The first calculation unit 2113 may calculate, for each registered fingerprint image, the partial similarity between one of the at least two target partial feature vectors and one of the at least two registered partial feature vectors for each of the at least two target partial feature vectors and each of the at least two registered partial feature vectors. In the example shown in Fig. 5, the first calculation unit 2113 may calculate the partial similarity between one of the four target partial feature vectors T1 to T4 and one of the four registered partial feature vectors R1 to R4 for each of the four target partial feature vectors T1 to T4 and each of the four registered partial feature vectors R1 to R4, thereby calculating 16 partial similarities. Specifically, for one registered fingerprint image RI, the first matching unit 211 may calculate the partial similarity between the first target partial feature vector T1 and the first registered partial feature vector R1, the partial similarity between the first target partial feature vector T1 and the second registered partial feature vector R2, the partial similarity between the first target partial feature vector T1 and the third registered partial feature vector R3, and the partial similarity between the first target partial feature vector T1 and the fourth registered partial feature vector R4. Furthermore, the first matching unit 211 may calculate the partial similarity between the second target partial feature vector T2 and the first registered partial feature vector R1, the partial similarity between the second target partial feature vector T2 and the second registered partial feature vector R2, the partial similarity between the second target partial feature vector T2 and the third registered partial feature vector R and R3, and the partial similarity between the second target partial feature vector T2 and the fourth registered partial feature vector R4. Furthermore, the first matching unit 211 may calculate the partial similarity between the third target partial feature vector T3 and the first registered partial feature vector R1, the partial similarity between the third target partial feature vector T3 and the second registered partial feature vector R2, the partial similarity between the third target partial feature vector T3 and the third registered partial feature vector R3, and the partial similarity between the third target partial feature vector T3 and the fourth registered partial feature vector R4. Furthermore, the first matching unit 211 may calculate the partial similarity between the fourth target partial feature vector T4 and the first registered partial feature vector R1, the partial similarity between the fourth target partial feature vector T4 and the second registered partial feature vector R2, the partial similarity between the fourth target partial feature vector T4 and the third registered partial feature vector R3, and the partial similarity between the fourth target partial feature vector T4 and the fourth registered partial feature vector R4.

[0044] The first calculation unit 2113 calculates a first similarity using the target partial feature vector and the registered partial feature vector for each of the multiple registered fingerprint images (step S26). The first calculation unit 2113 may use a representative partial similarity among the multiple partial similarities as the first similarity. The first calculation unit 2113 may calculate the first similarity based on the maximum partial similarity among the multiple partial similarities.

[0045] The first matching unit 211 determines whether the first similarity is equal to or greater than a first threshold (step S27). Alternatively, the first matching unit 211 may determine whether at least one of the 16 partial similarities is equal to or greater than a first threshold. In other words, the first matching unit 211 performs matching based on the similarity between partial regions of the fingerprint image. A method of extracting partial feature vectors from partial regions of the fingerprint image and matching fingerprint images using the similarity between the partial feature vectors is sometimes referred to as the first matching method.

[0046] If the first similarity is equal to or greater than the first threshold (step S27: Yes), the second matching unit 212 calculates the second similarity. That is, the second matching unit 212 may use registered fingerprint images whose first similarity is equal to or greater than the first threshold as the target for calculating the second similarity. For example, if the first similarity for 100 of 10,000 registered fingerprint images is equal to or greater than the first threshold, the second matching unit 212 may use those 100 registered fingerprint images as the target for calculating the second similarity.

[0047] If the first similarity is less than the first threshold value (step S27: No), the determination unit 213 determines that the target fingerprint image and the registered fingerprint image are not fingerprint images of the same individual.

[0048] The minutia extraction unit 2121 extracts information indicating minutiae (referred to as "target minutiae") TM as feature information from the entire target fingerprint image TI (step S28). A minutia may be a feature point such as an end point or a bifurcation point extracted from a fingerprint ridge. The information indicating a minutia may include information indicating the two-dimensional coordinates of the feature point and information indicating the directional angle of the ridge at the feature point.

[0049] The minutia storage unit 223 may include information RM (referred to as "registered minutia") indicating minutia as feature information extracted from the entire registered fingerprint image RI. The minutia storage unit 223 may include each of the registered minutia RM for each of the multiple registered fingerprint images RI.

[0050] The second matching unit 212 acquires the registered minutia RM from the minutia storage unit 223 (step S29). For example, if there are 100 registered fingerprint images to be used for calculating the second similarity, the second matching unit 212 may acquire, from the minutia storage unit 223, each of the registered minutia RM corresponding to the 100 registered fingerprint images RI.

[0051] The second calculation unit 2122 calculates a second similarity, which is the similarity between the target minutia TM and the registered minutia RM (step S30). In other words, the second calculation unit 2122 calculates the similarity between the entire fingerprint images. For example, the second calculation unit 2122 may align the center of the finger in the target fingerprint image with the center of the finger in the registered fingerprint image to calculate the similarity between the target minutia and the registered minutia. The second calculation unit 2122 may calculate each of the second similarities between the target minutia and each of the registered minutia.

[0052] In other words, the second matching unit 212 matches the fingerprint images based on the overall similarity of the fingerprint images. A method of extracting information indicating minutiae from a fingerprint image and matching the fingerprint images using the similarity between pieces of information indicating minutiae is sometimes referred to as the second matching method.

[0053] The determination unit 213 compares the target fingerprint image with the registered fingerprint image based on at least the second similarity (step S14). If at least the second similarity indicates that the target fingerprint image and the registered fingerprint image are more similar than a predetermined standard, the determination unit 213 may determine that the target fingerprint image and the registered fingerprint image are biometric images of the same individual if the second similarity is equal to or greater than a predetermined threshold. If the maximum second similarity among the multiple second similarities is equal to or greater than a predetermined threshold, the determination unit 213 may determine that the target fingerprint image and the registered fingerprint image corresponding to the maximum second similarity are biometric images of the same individual.

[0054] The determination unit 213 outputs the result of the comparison (step S15). The determination unit 213 may output information indicating whether the target fingerprint image and the registered fingerprint image are fingerprint images of the same individual. [2-3: Technical Effects]

[0055] In many cases, the first matching method is relatively fast, while the second matching method is relatively slow. The matching device 20 according to this disclosure first performs matching using the relatively fast first matching method, and only registers fingerprint images that have been successfully matched using the first matching method are used as matching targets for the second matching method. This reduces the number of matching targets for the second matching method. Reducing the number of matching targets for the second matching method is particularly useful when performing large-scale fingerprint authentication. The matching device 20 can achieve high-speed, high-precision fingerprint authentication. [3: Third Embodiment]

[0056] A third embodiment relating to a verification device, a verification method, and a recording medium will be described with reference to Figures 6 to 8. Below, the third embodiment relating to a verification device, a verification method, and a recording medium will be described using a verification device 30. Note that, for the third embodiment, descriptions that overlap with the descriptions of the first and second embodiments will be omitted as appropriate. Note that, in the drawings, parts common to the first and second embodiments are indicated by the same reference numerals.

[0057] 6, the calculation device 21 included in the matching device 30 includes, as logical functional blocks, a first matching unit 311, a second matching unit 312, and a determination unit 213. The first matching unit 311 may include a partial region acquisition unit 2111, a feature vector extraction unit 2112, and a first calculation unit 3113. The second matching unit 312 may include a minutiae extraction unit 2121, a second calculation unit 2122, and an alignment unit 3123. [3-1: Matching Method Executed by the Matching Device 30]

[0058] The collation method executed by the collation device 30 will be described with reference to Fig. 7. Fig. 7 is a flowchart showing an example of the flow of the collation executed by the collation device 30.

[0059] As shown in Fig. 7, the partial region acquisition unit 2111 acquires a target fingerprint image (step S21). The partial region acquisition unit 2111 acquires each of at least two target partial regions in the target fingerprint image (step S22). For example, as illustrated in Fig. 8, the partial region acquisition unit 2111 may divide the target fingerprint image TI into four parts and acquire each of the four target partial regions. Note that the target partial regions other than the lower right target partial region of the target fingerprint image TI in Fig. 8 (the upper left target partial region indicated by a circle, the upper right target partial region indicated by a triangle, and the lower left target partial region indicated by a square) are examples of partial regions in which no fingerprint is captured, partial regions in which a pattern cannot be recognized, etc., and do not mean that a circle, triangle, or square actually appears in the fingerprint image TI.

[0060] The feature vector extraction unit 2112 acquires partial feature information from each of the four target partial regions (step S23). The feature vector extraction unit 2112 may extract a target partial feature vector from each of the four target partial regions. In the example shown in FIG. 8 , the feature vector extraction unit 2112 may acquire a first target partial feature vector T1, a second target partial feature vector T2, a third target partial feature vector T3, and a fourth target partial feature vector T4. The first calculation unit 2113 acquires registered partial feature vectors from the partial feature vector storage unit 222 (step S24). In the example shown in FIG. 8 , the first calculation unit 2113 may acquire a first registered partial feature vector R1, a second registered partial feature vector R2, a third registered partial feature vector R3, and a fourth registered partial feature vector R4 from the partial feature vector storage unit 222.

[0061] The first calculation unit 3113 calculates the partial similarity between each of the target partial feature vectors and each of the registered partial feature vectors (step S25). In the example shown in Fig. 8, the first calculation unit 3113 may calculate the partial similarity between one of the four target partial feature vectors T1 to T4 and one of the four registered partial feature vectors R1 to R4 for each of the four target partial feature vectors T1 to T4 and each of the four registered partial feature vectors R1 to R4, thereby calculating 16 partial similarities.

[0062] The first calculation unit 3113 calculates the first similarity based on the multiple partial similarities (step S26). In the third embodiment, in step S26, the first calculation unit 3113 may determine the maximum partial similarity among the multiple partial similarities as the first similarity. Below, a case will be described in which the first calculation unit 3113 determines the maximum partial similarity among the multiple partial similarities as the first similarity. For example, in the example shown in FIG. 8 , the partial similarity between the third target partial feature vector T3 and the first registered partial feature vector R1 may be the maximum partial similarity, and the partial similarity between the third target partial feature vector T3 and the first registered partial feature vector R1 may be determined as the first similarity.

[0063] The first collation unit 311 determines whether the first similarity is equal to or greater than a first threshold (step S27). If the first similarity is equal to or greater than the first threshold (step S27: Yes), the alignment unit 3123 acquires, as an alignment reference, a combination of the target partial region corresponding to the first similarity and the registered partial region (step S31). In the example shown in Figure 8, the alignment unit 3123 may acquire, as an alignment reference, a combination of the target partial region corresponding to the third target partial feature vector T3 and the registered partial region corresponding to the first registered partial feature vector R1.

[0064] The alignment unit 3123 aligns the target partial region of the alignment reference with the registered partial region of the alignment reference (step S32). In the example shown in Fig. 8, the alignment unit 3123 may align the target partial region corresponding to the third target partial feature vector T3 with the registered partial region corresponding to the first registered partial feature vector R1.

[0065] The second calculation unit 3122 calculates the second similarity, which is the similarity between the target minutia and the registered minutia (step S30). In the third embodiment, the second calculation unit 3122 may calculate the second similarity in step S30 using the minutia aligned based on the partial region.

[0066] The determination unit 213 compares the target fingerprint image with the registered fingerprint image based on at least the second similarity (step S14). The determination unit 213 outputs the comparison result (step S15). [3-2: Technical Effects]

[0067] For example, there are cases where fingerprint images containing missing or misaligned fingerprints are to be compared. FIG. 8 illustrates an example of a case where fingerprints containing missing or misaligned fingerprints are compared. In the example illustrated in FIG. 8, the centers of the fingers in the target fingerprint image TI and the registered fingerprint image RI cannot be aligned. In particular, latent fingerprints may have missing or misaligned fingerprints, as in the target fingerprint image TI illustrated in FIG. 8. Note that FIG. 5 illustrates an example of a case where fingerprints without missing or misaligned fingerprints are compared. When there are large missing or misaligned fingerprints, fingerprint comparison may not be possible unless proper alignment is performed.

[0068] The matching device 30 according to this disclosure matches partial feature vectors extracted from partial regions of a fingerprint image, and can therefore determine which parts of a target fingerprint image correspond to which parts of a registered fingerprint image. Therefore, even when there is a large misalignment between the fingerprints included in the target fingerprint image and the registered fingerprint image, the matching device 30 can align and match the fingerprints. Even when there is a missing fingerprint and the target fingerprint image contains only a portion of the fingerprint, the matching device 30 can perform matching using the first matching method and then the second matching method to determine whether the target fingerprint image and the registered fingerprint image are fingerprint images of the same individual. The matching device 30 performs matching using the relatively fast first matching method and then the second matching method using the alignment standard obtained by matching using the first matching method, thereby enabling high-speed and accurate matching. [4: Fourth Embodiment]

[0069] A fourth embodiment relating to a verification device, a verification method, and a recording medium will be described with reference to FIG. 9 . Hereinafter, the fourth embodiment relating to a verification device, a verification method, and a recording medium will be described using a verification device 40. Note that, for the fourth embodiment, descriptions that overlap with the descriptions of the first to third embodiments will be omitted as appropriate. Note that, in the drawings, parts common to the first to third embodiments are designated by the same reference numerals.

[0070] 9 , the calculation device 21 included in the matching device 40 includes, as logical functional blocks, a first matching unit 311, a second matching unit 312, and a determination unit 413. The first matching unit 311 may include a partial region acquisition unit 2111, a feature vector extraction unit 2112, and a first calculation unit 3113. The second matching unit 312 may include a minutiae extraction unit 2121, a second calculation unit 2122, and an alignment unit 3123.

[0071] The determination unit 413 in the fourth embodiment calculates an integrated similarity by integrating the first similarity and the second similarity. The determination unit 413 may, for example, calculate an average value of the first similarity and the second similarity. The determination unit 413 may calculate the integrated similarity by weighting each of the first similarity and the second similarity. If the integrated similarity is equal to or greater than a second threshold, the determination unit 413 determines that the target biometric image and the registered biometric image are biometric images of the same individual. [Technical Effects]

[0072] The matching device 40 according to this disclosure can accurately identify an individual by using both the similarity of partial pattern features and the similarity of the variation of feature points in the pattern. This is a different approach from the first matching method and the second matching method. Therefore, by combining the first matching method and the second matching method, highly accurate fingerprint authentication can be achieved. [5: Supplementary Note]

[0073] Some or all of the above embodiments can be described as, but are not limited to, the following supplementary notes. [Supplementary Note 1] A matching device comprising: a first calculation means for calculating a first similarity using each of target partial feature information of at least two target partial regions in a target biometric image and each of registered partial feature information of at least two registered partial regions in a registered biometric image; a second calculation means for calculating a second similarity using target overall feature information of the entire target biometric image and registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold; and a determination means for outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity. [Supplementary Note 2] The matching device according to Supplementary Note 1, wherein the first calculation means calculates each of the first similarities using the target partial feature information and the registered partial feature information of each of the registered biometric images, and the second calculation means calculates a second similarity using the target overall feature information and the registered overall feature information for the registered biometric images for which the first similarity is equal to or greater than a first threshold. [Supplementary Note 3] The matching device according to Supplementary Note 1, wherein the first calculation means calculates the first similarity using the target portion feature information of each of the at least two target partial regions obtained by dividing the target biometric image and the registered portion feature information of each of the at least two registered partial regions obtained by dividing the registered biometric image. [Supplementary Note 4] The matching device according to Supplementary Note 1, wherein the target biometric image and the registered biometric image are each a biometric pattern image, and the target overall feature information and the registered overall feature information are each a minutia. [Supplementary Note 5] The matching device according to Supplementary Note 1, wherein the first calculation means calculates a similarity between one of the at least two pieces of target portion feature information and one of the at least two pieces of registered portion feature information for each of the at least two pieces of target portion feature information and each of the at least two pieces of registered portion feature information, and obtains the first similarity based on a plurality of the similarities.[Supplementary Note 6] The matching device according to Supplementary Note 5, wherein the first calculation means calculates the first similarity based on a maximum similarity among the plurality of similarities, defines a representative similarity among the plurality of similarities as the first similarity, and defines a maximum similarity among the plurality of similarities as the first similarity. [Supplementary Note 7] The matching device according to Supplementary Note 6, wherein the second calculation means, when the first similarity is equal to or greater than a first threshold, calculates the second similarity using the target overall feature information and the registered overall feature information, using as initial positions a position of one of the at least two target partial regions corresponding to the first similarity and a position of one of the at least two registered partial regions. [Supplementary Note 8] The matching device according to Supplementary Note 1, wherein the determination means calculates an integrated similarity by integrating the first similarity and the second similarity, and determines that the target biometric image and the registered biometric image are biometric images of the same individual when the integrated similarity is equal to or greater than a second threshold. [Supplementary Note 9] The matching device according to Supplementary Note 1, wherein the determination means determines that the target biometric image and the registered biometric image are not biometric images of the same individual when the first similarity is less than a first threshold. [Supplementary Note 10] The matching device according to Supplementary Note 1, further comprising: a storage means for storing the registered partial feature information for each of the registered partial regions and the registered overall feature information. [Supplementary Note 11] A matching method executed by a computer, comprising: calculating a first similarity using each of the target partial feature information for each of at least two target partial regions in the target biometric image and each of the registered partial feature information for each of at least two registered partial regions in the registered biometric image; when the first similarity is equal to or greater than a first threshold, calculating a second similarity using the target overall feature information of the entire target biometric image and the registered overall feature information of the entire registered biometric image; and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.[Supplementary Note 12] A recording medium having recorded thereon a computer program for causing a computer to execute a matching method, the method including: calculating a first similarity using each of target portion feature information of at least two target partial regions in a target biometric image and each of registered portion feature information of at least two registered partial regions in a registered biometric image; calculating a second similarity using target overall feature information of the entire target biometric image and registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold; and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

[0074] Furthermore, some or all of the configurations described in Supplementary Notes 2 to 10, which are dependent on Supplementary Notes 1 described above, and Supplementary Notes 11 and 12, may be dependent in the same manner as Supplementary Notes 2 to 10. Furthermore, not limited to Supplementary Notes 1, 11, and 12, some or all of the configurations described as Supplements may be dependent on various hardware, software, various recording means for recording software, or systems, within the scope of the above-described embodiments.

[0075] This disclosure may be modified as appropriate within the scope that does not contradict the gist or idea of ​​the invention that can be read from the claims and the entire specification, and the matching device, matching method, and program that involve such modifications are also included in the technical idea of ​​this disclosure.

[0076] 10, 20, 30, 40 Matching device 11, 2113, 3113 First calculation unit 12, 2122, 3122 Second calculation unit 13, 213, 413 Determination unit 211, 311 First matching unit 212, 312 Second matching unit 2111 Partial region acquisition unit 2112 Feature vector extraction unit 2121 Minutia extraction unit 222 Registered partial feature vector 223 Registered minutia 3123 Alignment unit

Claims

1. A matching device comprising: a first calculation means for calculating a first similarity using each of target partial feature information of at least two target partial regions in a target biometric image and each of registered partial feature information of at least two registered partial regions in a registered biometric image; a second calculation means for calculating a second similarity using target overall feature information of the entire target biometric image and registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold; and a determination means for outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

2. The matching device according to claim 1, wherein the first calculation means calculates each of the first similarities using the target partial feature information and the registered partial feature information of each of the plurality of registered biometric images, and the second calculation means calculates a second similarity using the target overall feature information and the registered overall feature information for the registered biometric images whose first similarity is equal to or greater than a first threshold.

3. The matching device described in claim 1, wherein the first calculation means calculates the first similarity using the target portion feature information of each of the at least two target partial areas into which the target biometric image is divided and the registered portion feature information of each of the at least two registered partial areas into which the registered biometric image is divided.

4. The matching device according to claim 1, wherein each of the target biometric image and the registered biometric image is a biometric pattern image, and each of the target overall feature information and the registered overall feature information is a minutia.

5. The matching device according to claim 1, wherein the first calculation means calculates the similarity between one of the at least two pieces of target partial feature information and one of the at least two pieces of registered partial feature information for each of the at least two pieces of target partial feature information and each of the at least two pieces of registered partial feature information, and determines the first similarity based on the multiple similarities.

6. The collation device according to claim 5, wherein the first calculation means determines the maximum similarity among the plurality of similarities as the first similarity.

7. The matching device described in claim 6, wherein the second calculation means, when the first similarity is equal to or greater than a first threshold, calculates the second similarity using the target overall feature information and the registered overall feature information, using as initial positions the position of one of the at least two target partial regions corresponding to the first similarity and the position of one of the at least two registered partial regions.

8. The matching device according to claim 1, wherein the determination means calculates an integrated similarity by integrating the first similarity and the second similarity, and determines that the target biometric image and the registered biometric image are biometric images of the same individual if the integrated similarity is equal to or greater than a second threshold value.

9. The matching device according to claim 1, wherein the determination means determines that the target biometric image and the registered biometric image are not biometric images of the same individual when the first similarity is less than a first threshold value.

10. The matching device according to claim 1, further comprising a storage means for storing the registered partial feature information for each registered partial region and the registered overall feature information.

11. A matching method executed by a computer, comprising: calculating a first similarity using each of target portion feature information of at least two target partial regions in a target biometric image and each of registered portion feature information of at least two registered partial regions in a registered biometric image; if the first similarity is equal to or greater than a first threshold, calculating a second similarity using target overall feature information of the entire target biometric image and registered overall feature information of the entire registered biometric image; and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

12. A recording medium having recorded thereon a computer program for causing a computer to execute a matching method, the method comprising: calculating a first similarity using each of target portion feature information of at least two target partial regions in a target biometric image and each of registered portion feature information of at least two registered partial regions in a registered biometric image; calculating a second similarity using target overall feature information of the entire target biometric image and registered overall feature information of the entire registered biometric image when the first similarity is equal to or greater than a first threshold; and outputting a matching result between the target biometric image and the registered biometric image based on at least the second similarity.

Images