Personal authentication program, personal authentication method, and personal authentication device

Abstract

The person authentication program is executed by a computer connected to a first camera that captures an entire image of one or more persons located in a region to be monitored, and a second camera that captures an entire image of one or more persons located in a region to be monitored. And a person authentication program for causing the computer to execute a process for authenticating the person in front of the second camera when the person is recognized to be present in front of the second camera, and for capturing an image of the head of the person when the person is authenticated at the entrance of the area to be monitored. Selecting a specific head image satisfying a condition from among head images representing the head of the person appearing in the captured image of the first camera; the specific head image is compared with an image captured by the second camera, and an identifier obtained as an authentication result is associated with a specific person corresponding to the specific head image for which the similarity is calculated to be equal to or greater than a threshold similarity.

Description

Technical Field

[0001] This case involves personal authentication procedures, personal authentication methods, and personal authentication devices. Background Technology

[0002] To restrict the number of people passing through entrances and exits, sometimes it is necessary to take a picture of the person's face from the front for authentication and control the opening and closing of the entrance or exit based on the authentication result. In this case, in order to monitor unauthorized persons who want to pass through the entrance or exit without authentication, it is sometimes necessary to take a picture from above of the area around the entrance or exit for tracking of the persons passing through. In addition, it is known that there are person detection devices that detect persons in each image of images taken by multiple cameras and establish corresponding detection of the same person (see, for example, Patent Document 1).

[0003] In addition, there are known technologies such as image tracking technology and technology that automatically determines the timing of shooting by recognizing a smiling face from the subject included in the image (see, for example, Patent Documents 2-5).

[0004] Patent Document 1: Japanese Patent Application Publication No. 2015-114917

[0005] Patent Document 2: Japanese Patent Application Publication No. 2010-016878

[0006] Patent Document 3: Japanese Patent Application Publication No. 2000-222576

[0007] Patent Document 4: U.S. Patent Application Publication No. 2017 / 0178345

[0008] Patent Document 5: U.S. Patent Application Publication No. 2020 / 0074816

[0009] However, when authenticating individuals by photographing their faces from the front, including those passing through, multiple individuals may accidentally appear in the authentication process. In such cases, even if people are tracked by photographing them from above near the entrance / exit, there is a risk that it may be impossible to uniquely identify which person is the target for authentication. For example, a scenario where a limit on the number of people being authenticated is set near the entrance / exit has been considered, but the area around the entrance / exit is likely to be chaotic, potentially reducing the convenience of the authentication process.

[0010] Furthermore, a scenario was envisioned where specific patterns, such as smiley faces, could be identified in an image taken from above, and the person in question could be identified as the target of the authentication process, thus enabling tracking. However, if the pattern differentiation accuracy is low, it may not be able to uniquely identify the target person for authentication, preventing normal tracking processing. Summary of the Invention

[0011] Therefore, in one respect, its purpose is to provide a person authentication procedure, person authentication method, and person authentication device that uniquely identifies the person to be authenticated.

[0012] In one embodiment, the person authentication procedure is a person authentication procedure executed by a computer connected to a first camera and a second camera. The first camera takes a full picture of one or more persons in the monitored area, and the second camera takes a picture of the head of the persons when authenticating them at the entrance of the monitored area. The person authentication procedure is used to cause the computer to perform the following processes: when it is identified that a person is in front of the second camera, select a specific head image that meets certain conditions from head images representing the head of the person appearing in the image captured by the first camera; and compare the specific head image with the image captured by the second camera, and establish a correspondence between an identifier obtained as an authentication result and a specific person corresponding to the specific head image whose similarity is calculated to be above a threshold similarity.

[0013] It can uniquely identify the person being authenticated. Attached Figure Description

[0014] Figure 1 This is an example of a personnel surveillance system.

[0015] Figure 2 This is an example of a side view near an entrance in the monitored area.

[0016] Figure 3 This is an example of the hardware structure of a person authentication device.

[0017] Figure 4 This is an example of the functional structure of a person authentication device.

[0018] Figure 5 (a) is an example of authentication information. Figure 5 (b) is another example of authentication information.

[0019] Figure 6 (a) is an example of a wide-area image showing two figures. Figure 6 (b) is an example of a wide-area image used for person detection and head detection.

[0020] Figure 7 This is an example of the second header information.

[0021] Figure 8 This is a flowchart illustrating an example of the processing performed by the person authentication device.

[0022] Figure 9 This is another example of second header information.

[0023] Figure 10 This is an example of generating the first header information.

[0024] Figure 11 This is an example of the best choice for a head image.

[0025] Figure 12 (a) is an example of similarity calculation. Figure 12 (b) is an example of the creation of my ID.

[0026] Figure 13 This is an example of a wide-area image of a person.

[0027] Figure 14 (a) is an example of a wide-area image where the distance between the person and the authentication camera is above a threshold distance. Figure 14 (b) is an example of a wide-area image illustrating the distances between two people and the authentication camera. Detailed Implementation

[0028] The following description, with reference to the accompanying drawings, illustrates the method of implementing this case.

[0029] like Figure 1 As shown, the person monitoring system ST includes a surveillance camera 10, an authentication camera 20, an authentication information reader 21, and a person authentication device 100. The surveillance camera 10 is an example of a first camera, and the authentication camera 20 is an example of a second camera. The person authentication device 100 is an example of a computer. The person monitoring system ST may also include multiple surveillance cameras 11, 12, 13, 14, and 15.

[0030] Surveillance cameras 10, 11, 12, 13, 14, and 15, authentication camera 20, and authentication information reader 21 are connected to the person authentication device 100 via a communication network NW. The communication network NW includes at least one of a LAN (Local Area Network), a WAN (Wide Area Network), and the Internet.

[0031] The person authentication device 100 is connected to an input device 710 and a display device 720. The input device 710 may include, for example, a keyboard, mouse, or touch panel. The display device 720 may include, for example, a liquid crystal display (LCD). The input device 710 is operated, for example, by the administrator P0 of the person monitoring system ST. Based on the operation of the input device 710, the person authentication device 100 displays, for example, a wide-area image (hereinafter referred to as a wide-area image) captured by the surveillance camera 10 on the display device 720. Furthermore, based on the wide-area images captured by at least one of the surveillance cameras 10, 11, 12, 13, 14, and 15, the person authentication device 100 establishes a corresponding tracking mechanism for the same person P1 appearing in the wide-area image.

[0032] Surveillance cameras 10, 11, 12, 13, 14, and 15 all capture overhead, color images of the monitored area AR. For example, if the monitored area AR is an office, surveillance cameras 10, 11, 12, 13, 14, and 15 are all positioned on the ceiling, above the heads of individuals P1 and P2. Thus, for example, surveillance camera 10 captures the entire area, including the full bodies of individuals P1 and P2 within the monitored area AR, authentication camera 20, authentication information reader 21, etc. Furthermore, the monitored area AR is not limited to offices; it can also be medical facilities such as hospitals or clinics, public facilities such as airports or train stations, or commercial facilities such as shopping malls.

[0033] Furthermore, the surveillance camera 10 is configured to capture images of person P1 located in front of the authentication camera 20 and the authentication information reader 21. As a result, person P1 appears partially or entirely within the wide-area image of the surveillance camera 10. The surveillance camera 10 is positioned such that person P1's head does not overlap with the authentication camera 20 and the authentication information reader 21 within the wide-area image. Therefore, person P1's face is not hidden by the authentication camera 20 and the authentication information reader 21 within the wide-area image of the surveillance camera 10.

[0034] The certified camera 20, for example, takes a color photograph of the head of person P1 from the front. Specifically, as... Figure 2 As shown, the authentication camera 20 is mounted on top of the base 22 via a tripod. Furthermore, the shooting range of the authentication camera 20 is adjusted to shoot the head Ph of the person P1 from the front. Additionally, Figure 2 Corresponding to Figure 1 The designated area ARy is shown. The tripod can also be extended to remove the base 22, or the tripod can be removed from the authentication camera 20 and the authentication camera 20 can be directly mounted on the top surface of the base 22.

[0035] On the other hand, such as Figure 1As shown, the authentication information reader 21 authenticates person P1, for example, at the entrance ARx of the monitored object area AR. Figure 2 As shown, the authentication information reader 21 is disposed on the top surface of the pedestal 22. Additionally, the authentication information reader 21 is disposed in front of the authentication camera 20. The authentication information reader 21 can authenticate person P1 by reading biometric information related to biometric authentication and communicating with the person authentication device 100. The authentication information reader 21 can also authenticate person P1 by reading card information (such as personal identification number) recorded on the ID (Identifier) ​​card carried by person P1 and communicating with the person authentication device 100. The ID card is an example of a person identification medium. Thus, the authentication information reader 21 authenticates person P1, and the authentication camera 20 supplements the authentication process based on the authentication information reader 21 by taking a picture of person P1's head Ph.

[0036] Furthermore, the aforementioned biometric authentication can also be fingerprint authentication, vein authentication, or iris authentication. In this embodiment, the case where authentication of person P1 is performed using both the authentication camera 20 and the authentication information reader 21 is described. On the other hand, when a facial authentication camera is used as the authentication camera 20, the authentication information reader 21 can be excluded from the person monitoring system ST. The facial authentication camera can authenticate person P1 using a facial image (i.e., an image of a face) by communicating with the person authentication device 100.

[0037] Thus, when person P1 enters the monitored area AR, the authentication information reader 21 reads person P1's own biometric information or the card information of the ID card carried by person P1. The authentication information reader 21 is positioned in front of the authentication camera 20, so when the authentication information reader 21 authenticates person P1, the authentication camera 20 captures a picture of person P1's head Ph. The biometric information or card information is sent from the authentication information reader 21 to the person authentication device 100. The headshot image (hereinafter referred to as the authentication head image) including person P1's head Ph is sent from the authentication camera 20 to the person authentication device 100.

[0038] The person authentication device 100 authenticates person P1 based on biometric information or card information, and maintains the authentication result and the authenticated head image of person P1 as first head information in a correlated manner. Details will be described later, but if the person authentication device 100 receives the authenticated head image of person P1, it infers the direction of the head Ph based on the authenticated head image. Specifically, the person authentication device 100 calculates a head orientation vector based on the authenticated head image. The head orientation vector, for example, represents the orientation of person P1's face. The person authentication device 100 also maintains this head orientation vector in a correlated manner with the authentication result, etc., as first head information.

[0039] On the other hand, the surveillance camera 10 captures images of the authentication information reader 21 authenticating person P1. The wide-area image from the surveillance camera 10 is periodically (e.g., every few seconds) transmitted from the surveillance camera 10 to the person authentication device 100. Figure 1 As shown, if the shooting range of the surveillance camera 10 includes not only person P1 but also person P2, then the full body of person P1 and the full body of person P2 will appear within the shooting range of the surveillance camera 10.

[0040] The details will be described later. However, if the person authentication device 100 receives a wide-area image from the surveillance camera 10, it detects the full-body images of persons P1 and P2 appearing in the wide-area image, and assigns different temporary identification information (hereinafter referred to as temporary ID) to persons P1 and P2 respectively. In addition, the person authentication device 100 detects the head from the full-body images of persons P1 and P2, and infers the head orientation based on the head image representing the detected head (hereinafter referred to as the detected head image). That is, the person authentication device 100 calculates the head orientation vector.

[0041] If the person authentication device 100 calculates the head direction vector, it stores the temporary ID, head direction vector, detected head image, and the acquisition date and time of the wide-area image in a related manner as second head information. If the person authentication device 100 stores the second head information, it determines the same person based on the first and second head information. That is, the person authentication device 100 determines person P1, who is in front of the authentication camera 20, as the person to be authenticated, from persons P1 and P2 appearing in the wide-area image of the surveillance camera 10. If the person authentication device 100 determines the same person, it replaces the temporary ID assigned to person P1 with a personal ID that identifies person P1.

[0042] In this way, even without detecting specific patterns such as smiley faces, the person authentication device 100 can accurately establish a correspondence between the person's ID and the person P1 who has been assigned a temporary ID. Then, it can track the movement of person P1 within the monitored area AR. Therefore, for example, even if person P1 commits illegal acts within the monitored area AR, the person authentication device 100 can uniquely identify person P1. Furthermore, the person authentication device 100 can uniquely grasp the movement of person P1 within the monitored area AR.

[0043] Next, refer to Figure 3 The hardware structure of the character authentication device 100 will be explained.

[0044] The character authentication device 100 includes a CPU (Central Processing Unit) 100A as a processor, RAM (Random Access Memory) 100B as memory, and ROM (Read-Only Memory) 100C. RAM 100B includes DRAM (Dynamic RAM) and SRAM (Static RAM). SRAM may also be included in the CPU 100A. The character authentication device 100 includes a network I / F (interface) 100D and an HDD (Hard Disk Drive) 100E. Alternatively, an SSD (Solid State Drive) may be used instead of the HDD (Hard Disk Drive) 100E.

[0045] The character authentication device 100 may also include at least one of the following, as needed: input I / F 100F, output I / F 100G, input / output I / F 100H, and drive device 100I. The CPU 100A and drive device 100I are interconnected via an internal bus 100J. That is, the character authentication device 100 can be implemented by a computer.

[0046] An input device 710 is connected to input I / F100F. A display device 720 is connected to output I / F100G. A semiconductor memory 730 is connected to input / output I / F100H. The semiconductor memory 730 may be, for example, a USB (Universal Serial Bus) memory, flash memory, etc. Input / output I / F100H reads the character authentication program stored in the semiconductor memory 730. Input I / F100F and input / output I / F100H may have, for example, a USB port. Output I / F100G may have, for example, a display port.

[0047] A portable recording medium 740 is inserted into the drive unit 100I. The portable recording medium 740 may be a removable disk such as a CD (Compact Disc)-ROM or a DVD (Digital Versatile Disc). The drive unit 100I reads the person authentication program recorded on the portable recording medium 740. The network I / F 100D may include, for example, a LAN port and communication circuitry. The communication circuitry may include either or both of wired and wireless communication circuitry. The network I / F 100D is connected to a communication network NW.

[0048] The person authentication program stored in at least one of ROM 100C, HDD 100E, and semiconductor memory 730 is temporarily stored in RAM 100B by CPU 100A. The person authentication program recorded on portable recording medium 740 is also temporarily stored in RAM 100B by CPU 100A. CPU 100A executes the stored person authentication program, thereby performing the various functions described later, and also executing the person authentication method including the various processes described later. Furthermore, the person authentication program can be set to correspond to the flowchart described later.

[0049] Reference Figures 4-7 The functional structure of the person authentication device 100 will be explained. Furthermore, in... Figure 4 The image shows the main functional components of the person authentication device 100.

[0050] like Figure 4 As shown, the person authentication device 100 includes a storage unit 110, a processing unit 120, an input unit 130, an output unit 140, and a communication unit 150. The storage unit 110 can be implemented using either or both of the RAM 100B and HDD 100E described above. The processing unit 120 can be implemented using the CPU 100A described above. The input unit 130 can be implemented using the input I / F 100F described above. The output unit 140 can be implemented using the output I / F 100G described above. The communication unit 150 can be implemented using the network I / F 100D described above.

[0051] The storage unit 110, processing unit 120, input unit 130, output unit 140, and communication unit 150 are interconnected. The storage unit 110 contains authentication information DB (Data Base) 111 and header information DB 112. The processing unit 120 contains a header information generation unit 121 and a person identification unit 122. The person identification unit 122 is an example of selecting a unit and establishing a corresponding unit.

[0052] The authentication information DB111 pre-stores the authentication information required for authenticating multiple characters, including characters P1 and P2. For example... Figure 5 As shown in (a), the authentication information includes multiple items such as authentication ID, authentication type, biometric information, card information, and person ID. The authentication ID item contains an identifier that identifies the authentication information. The authentication type item contains the authentication method. Authentication methods include, for example, biometric authentication and card authentication.

[0053] The Biometric Information section contains biometric information that uniquely identifies each of the multiple individuals. When the authentication method is biometric authentication, biometric information is also recorded in the Biometric Information section. The Card Information section contains card information recorded on the ID cards carried by each of the multiple individuals. When the authentication method is card authentication, card information is also recorded in the Card Information section. The Person ID section contains an identifier that identifies the multiple individuals. For example, if biometric information "#%#%#%" is recorded as the biometric information of person P1 in the Biometric Information section, then based on this biometric information "#%#%#%", the person ID "tanaka" is uniquely determined as the person ID of person P1.

[0054] Furthermore, in the case where a facial recognition camera is used as the recognition camera 20, such as Figure 5 As shown in (b), the authentication information can include multiple items such as authentication ID, facial image, and person ID. For example, if a facial image with a pentagonal outline is registered as the facial image of person P1 in the facial image item, the person ID "tanaka" can be uniquely determined as the person ID of person P1 based on this facial image. Alternatively, facial image-based authentication can utilize, for example, hair color.

[0055] The header information generation unit 121 generates the first header information and the second header information described above. If the header information generation unit 121 generates the first header information, it retains the first header information. If the header information generation unit 121 generates the second header information, it saves the second header information in the header information DB112.

[0056] To explain in more detail, the head information generation unit 121 periodically acquires wide-area images from the surveillance camera 10. For example, as Figure 6 As shown in (a), if a wide-area image is acquired, the head information generation unit 121 is based on a known person detection technique using OpenCV, such as... Figure 6 As shown in (b), person detection is performed on all persons P1 and P2 appearing in the wide-area image. If the head information generation unit 121 performs person detection, it assigns a temporary ID to each of persons P1 and P2. For example, the head information generation unit 121 assigns a temporary ID "1" to person P1 and a temporary ID "2" to person P2.

[0057] If a temporary ID is assigned, the head information generation unit 121 performs face detection (or face recognition) on each of persons P1 and P2 based on known face detection (or face recognition) techniques using OpenCV, and obtains the detected head image. If the head information generation unit 121 obtains the detected head image, it calculates the head orientation vector of each of persons P1 and P2 based on known head orientation inference techniques (or face orientation inference techniques) such as HOPE-Net and FaceRig and the detected head image.

[0058] The head orientation inference technique includes, for example, the following processing: detecting six coordinates of landmarks such as the corners of the eyes, the tip of the nose, and the corners of the mouth from the detected head image, and calculating the virtual position of the camera based on the detected coordinates, thereby inferring the head orientation. The head orientation (or facial orientation) of each person P1 and P2 is inferred from the head orientation vector calculated by the head information generation unit 121.

[0059] Thus, if the header information generation unit 121 calculates the header direction vector, it saves the second header information, which is formed by associating the temporary ID, the header direction vector, the detected header image, and the acquisition date and time of the wide-area image, in the header information DB112. Therefore, as... Figure 7 As shown, the header information DB112 stores the second header information of characters P1 and P2 respectively.

[0060] The second header information includes several items such as a temporary ID, a head orientation vector, a head image, and the date and time of acquisition. The temporary ID is registered in the temporary ID field. The head orientation vector is registered in the head orientation vector field. The head orientation vector is represented, for example, by roll angle "r4", pitch angle "p4", yaw angle "y4", etc. The detected head image is registered in the head image field. The date and time of acquisition is registered in the date and time of acquisition field. The date and time of acquisition is represented, for example, in the form YYYYMMDDHHMMSS. Furthermore, the header information generation unit 121 generates the first header information in essentially the same way as the second header information, except for acquiring the date and time; the details of the generation of the first header information will be described later.

[0061] The person identification unit 122 identifies the same person based on the first head information and the second head information. More specifically, the person identification unit 122 determines whether each detected head image in the detected head images included in the second head information meets predetermined conditions, and selects a specific detected head image that meets the predetermined conditions. The predetermined conditions are, for example, the relationship between the head direction vectors included in the first head information and the head direction vectors included in the second head information. That is, the predetermined conditions are related to the angle of the head direction. When the head direction vectors are in the closest possible relationship to each other, that is, when the angles of the head directions are in the closest possible relationship, the person identification unit 122 determines that the predetermined conditions are met.

[0062] In this case, the person identification unit 122 selects the detected head image that is most similar to the head direction vector included in the first head information for each temporary ID as the specific best head image. Moreover, the person identification unit 122 compares the best head image of each temporary ID with the certified head image based on the certification camera 20, and establishes a correspondence between the person ID obtained as the certification result and any one of the specific persons P1 and P2 corresponding to the best head image with a similarity of more than a threshold.

[0063] Therefore, even if persons P1 and P2 appear in a wide-area image, the person identification unit 122 can, for example, accurately identify person P1 as the person to be authenticated. Then, the person identification unit 122 can track person P1 based on the wide-area images from multiple surveillance cameras 10, 11, 12, 13, 14, and 15. Furthermore, the person identification unit 122 can also select, for each temporary ID, a detected head image that is correlated with the head direction vector most closely related to the head direction vector included in the first head information and the second most closely related head direction vector as a specific optimal head image.

[0064] Next, refer to Figures 8 to 14 An example of the operation of the character authentication device 100 will be explained.

[0065] First, such as Figure 8As shown, the head information generation unit 121 acquires a wide-area image from the surveillance camera 10 (step S1). If a wide-area image is acquired, the head information generation unit 121 performs person detection (step S2) and saves the second head information in the head information DB112 (step S3). If the second head information is saved, the head information generation unit 121 determines whether the person has been authenticated (step S4). For example, if the person P1 does not allow the authentication information reader 21 to read the person P1's own biometric information or the person P1's card information, the head information generation unit 121 determines that there is no person authentication (step S4: No). Thus, the head information generation unit 121 repeats the processing of steps S1 to S3.

[0066] The result is, as Figure 9 As shown, the header information DB112 stores multiple pieces of second header information while appending them in units of a few seconds. Furthermore, the header information generation unit 121 can also store the second header information in the header information DB112 until a predetermined storage period (e.g., a few minutes), and delete the second header information that has expired from the header information DB112. This suppresses the reduction of the storage capacity of the header information DB112.

[0067] Here, for example, if person P1 performs an authentication action that causes authentication information reader 21 to read person P1's own biological information, then the header information generation unit 121 determines that there is authentication (step S4: Yes). In the case of authentication, the header information generation unit 121 obtains person P1's biological information from authentication information reader 21. Therefore, the header information generation unit 121 can generate the header information based on the biological information and the authentication information stored in authentication information DB111 (see...). Figure 5 (a) of, such as Figure 10 As shown in the upper part, the person ID "tanaka" of person P1 is determined as the authentication result. If the authentication result is determined, the header information generation unit 121 determines whether more than two people have appeared in the recent wide-area image (step S5).

[0068] For example, by the presence of figures P1 and P2 in a wide-area image (refer to...) Figure 6 (a) , and in the case where two or more people appear in the wide-area image (step S5: Yes), the head information generation unit 121 acquires the authentication head image from the authentication camera 20 (step S6). That is, at a position less than 1m (meters) in front of the authentication camera 20, the person P1 who performed the authentication action for the authentication information reader 21 is facing forward and standing still. Therefore, as Figure 10 As shown in the middle part, the head information generation unit 121 can acquire the authentication head image captured by the authentication camera 20 with high precision.

[0069] If an authentication header image is obtained, the header information generation unit 121 generates first header information (step S7) and retains it. That is, as... Figure 10 As shown in the lower part, the head information generation unit 121 generates and maintains first head information that associates the person ID "tanaka" determined as the authentication result, the authentication head image, and the head direction vector calculated based on the authentication head image.

[0070] Furthermore, the first header information includes several items such as the user ID, head direction vector, and head image. The user ID item records the user ID determined as the authentication result. The head direction vector item records the head direction vector calculated based on the authenticated head image. The head image item records the authenticated head image. Thus, the first header information differs from the second header information in that it does not include the date and time of acquisition.

[0071] If the head information generation unit 121 generates and maintains the first head information, the person identification unit 122 selects the best head image (step S8). More specifically, as... Figure 11 As shown, the person identification unit 122 calculates the vector difference between the head direction vector included in the first head information and the head direction vector included in the second head information for each angle. Therefore, it calculates the vector differences for roll angle, pitch angle, and yaw angle. If a vector difference is calculated, the person identification unit 122 squares the vector differences for each angle and calculates the sum. For each temporary ID, it selects the head image associated with the head direction vector that has the smallest sum as the specific optimal head image. That is, the person identification unit 122 determines the head direction vector from the second head information that is closest to the head direction vector included in the first head information based on the least squares method. Furthermore, the person identification unit 122 selects the detected head image associated with the determined head direction vector for each temporary ID as the specific optimal head image.

[0072] If this implementation method is used, then as follows: Figure 11 As shown, the person identification unit 122 selects the detected head image, which includes the temporary ID "1" and the head direction vectors "r4, p4, y4", as the specific optimal head image. Additionally, the person identification unit 122 selects the detected head image, which includes the temporary ID "2" and the head direction vectors "r5, p5, y5", as the specific optimal head image.

[0073] If the best head image is selected, the person identification unit 122 calculates the similarity (step S9). More specifically, as... Figure 12As shown in (a), the person identification unit 122 calculates the image feature quantities of the certified head image 50 included in the first head information, and the image feature quantities of the two selected best head images 51 and 52, and calculates the similarity between the certified head image 50 and the best head images 51 and 52 based on the image feature quantities. For example, the person identification unit 122 can calculate the image feature quantities and the similarity based on the hair color, facial contour, etc. of the heads belonging to the certified head image 50 and the best head images 51 and 52.

[0074] In this embodiment, the person identification unit 122 calculates 98% similarity between the authenticated head image 50 included in the first head information and the best head image 51 associated with the temporary ID "1". Additionally, the person identification unit 122 calculates 12% similarity between the authenticated head image 50 included in the first head information and the best head image 52 associated with the temporary ID "2". Thus, by pre-reducing the number of best head images of the objects to be similared using the head direction vector, the person identification unit 122 can complete the similarity calculation in a shorter time compared to calculating all similarities.

[0075] If a similarity is calculated, the person identification unit 122 determines whether the similarity is above a threshold (step S10). More specifically, the person identification unit 122 determines whether the similarity for each best head image is above a threshold similarity. The threshold similarity can be set appropriately in advance; in this embodiment, for example, 90% is set as the threshold similarity. If all similarities are less than the threshold (step S10: No), the head information generation unit 121 returns to the processing in step S1.

[0076] On the other hand, if any similarity score exceeds the threshold (step S10: Yes), the person identification unit 122 establishes a corresponding ID for the person (step S11), and the process ends. More specifically, as... Figure 12 As shown in (b), the person identification unit 122 establishes a correspondence between the person ID and the best head image 51 by replacing the temporary ID "1" associated with the best head image 51, which has a similarity of more than a threshold, with the person ID "tanaka". Similarly, it establishes a correspondence between the person ID and the best head image 51 by replacing the temporary ID "1" associated with the person P1 appearing in the wide-area image with the person ID "tanaka". In this way, the person identification unit 122 can identify both the person P1 appearing in the wide-area image and the person P1 being authenticated.

[0077] Furthermore, in the processing of step S5, if no more than two people appear in the wide-area image (step S5: No), the person identification unit 122 executes the processing of step S10 and ends the processing. For example, as Figure 13 As shown, when only one person P1 appears in the wide-area image, since no person P2 appears in the wide-area image, the person identification unit 122 can uniquely identify person P1 as the person to whom the corresponding ID is established. Thus, when no more than two people appear in the wide-area image, the head information generation unit 121 and the person identification unit 122 can avoid the processing steps S6 to S10. That is, when no more than two people appear in the wide-area image, the person identification unit 122 can suppress the processing load and establish the corresponding ID for the person in a short time.

[0078] In addition, such as Figure 14 As shown in (a), when the authentication camera 20 is a facial authentication camera, if the distance between person P1 and the authentication camera 20 is, for example, more than 1 meter, the authentication camera 20 may not be able to capture a high-precision image of person P1's authenticated head. When using the authentication information reader 21, person P1 is close to the authentication information reader 21, so the distance between person P1 and the authentication camera 20 is more likely to be less than 1 meter. However, when using a facial authentication camera, it is also possible that the distance between person P1 and the authentication camera 20 is more than 1 meter, and a high-precision image of person P1's authenticated head may not be captured.

[0079] Therefore, when using a facial recognition camera, if more than two people appear in the wide-area image during the processing in step S5, then immediately following that processing, such as... Figure 14 As shown in (b), the head information generation unit 121 determines whether the distances between persons P1 and P2 and the authentication camera 20 are less than a threshold distance. The head information generation unit 121 can determine whether the distances between persons P1 and P2 and the authentication camera 20 are less than the threshold distance based on multiple objects included in the wide-area image. The threshold distance can be several meters, such as 1 meter, or tens of centimeters.

[0080] If the distance between person P1 and the authentication camera 20, and the distance between person P2 and the authentication camera 20, are both above the threshold distance, the head information generation unit 121 determines that there is no person to be authenticated and returns to the processing in step S1. Therefore, the head information generation unit 121 and the person identification unit 122 can avoid the processing in steps S6 to S10.

[0081] On the other hand, if the distance between person P1 and authentication camera 20 is less than a threshold distance, and the distance between person P2 and authentication camera 20 is greater than or equal to the threshold distance, person identification unit 122 can uniquely determine person P1 as the person to be identified for authentication. In this case, head information generation unit 121 and person identification unit 122 can also avoid the processing in steps S6 to S10.

[0082] When both the distance between person P1 and the authentication camera 20 and the distance between person P2 and the authentication camera 20 are less than the threshold distance, the head information generation unit 121 and the person identification unit 122 can proceed to the processing of step S10 after performing the processing of steps S6 to S9.

[0083] The preferred embodiments of the present invention have been described in detail above, but are not limited to the specific embodiments of the present invention. Various modifications and alterations can be made within the scope of the spirit of the present invention as described in the claims.

[0084] For example, the person identification unit 122 determines whether to select a specific optimal head image that meets the aforementioned conditions based on the correlation between the facilities of the monitored target area AR and the types of clothing commonly used in those facilities. Furthermore, if the correlation is higher than a threshold correlation, the person identification unit 122 avoids selecting that specific optimal head image that meets the prescribed conditions. For example, if a person wearing a white coat commonly used in facilities such as hospitals appears in a wide-area image, the likelihood of them committing illegal acts is considered low. Therefore, the person identification unit 122 excludes such a person from the tracked targets, thereby limiting the tracked targets to a subset of people, reducing processing load, and effectively identifying people.

[0085] Explanation of reference numerals in the attached figures: ST…person monitoring system; AR…monitored object area; P1, P2…person; 10, 11, 12, 13, 14, 15…monitoring cameras; 20…authentication camera; 21…authentication information reader; 100…person authentication device; 110…storage unit; 111…authentication information database; 112…head information database; 120…processing unit; 121…head information generation unit; 122…person identification unit.

Claims

1. A person authentication program, executed by a computer connected to a first camera and a second camera, wherein the first camera takes a full-body image of one or more persons within a monitored area, and the second camera takes an image of the head of each person when authenticating them at the entrance of the monitored area, wherein... The person authentication program is used to enable the computer to perform the following processes: If it is determined that the person is in front of the second camera, a specific head image that meets the conditions is selected from the head images representing the head of the person appearing in the image captured by the first camera. as well as The specific head image is compared with the image captured by the second camera, and the identifier obtained as the authentication result is used to establish a correspondence with the specific person corresponding to the specific head image that has a similarity of more than a threshold.

2. The personal authentication procedure according to claim 1, characterized in that, In the process of selecting the specific head image, if it is identified that the person is in front of the second camera based on the image captured by the first camera, the specific head image whose head direction is similar to that of the head image captured by the second camera is selected from the head images representing the head of the person appearing in the image captured by the first camera.

3. The personal authentication procedure according to claim 1 or 2, characterized in that, In the process of selecting the specific head image, if it is identified that the person is in front of the second camera based on the image captured by the first camera, the head of the person appearing in the image captured by the first camera is detected, the information is stored in the database until the information has expired, and the information that has expired is deleted from the database. The information is obtained by establishing a correspondence between the head image representing the detected head, the head direction inferred from the head image, and a temporary identifier for identifying the person corresponding to the head image per unit time.

4. The personal authentication procedure according to claim 1 or 2, characterized in that, In the process of selecting the specific head image, if the person identified as being in front of the second camera based on the image captured by the first camera is a single person, the selection of the specific head image that meets the condition is avoided.

5. The personal authentication procedure according to claim 1 or 2, characterized in that, In the process of selecting the specific head image, if the person is identified as being located at a position above a separation threshold distance from the second camera based on the image captured by the first camera, the selection of the specific head image that meets the condition is avoided.

6. The personal authentication procedure according to claim 1 or 2, characterized in that, In the process of selecting the specific head image, it is determined whether to select the specific head image that meets the conditions based on the correlation between the facilities in the monitored area and the types of clothing commonly used in the facilities, and if the correlation is higher than a threshold correlation, the selection of the specific head image that meets the conditions is avoided.

7. The personal authentication procedure according to claim 1 or 2, characterized in that, The second camera is a facial authentication camera that authenticates the face of the person being identified.

8. The personal authentication procedure according to claim 1 or 2, characterized in that, The second camera is positioned within the shooting range of the first camera.

9. A person authentication method, executed by a computer connected to a first camera and a second camera, wherein the first camera takes a full-body photo of one or more persons located in a monitored area, and the second camera takes a photo of the head of each person when authenticating them at the entrance of the monitored area, wherein... In the aforementioned person authentication method, the computer performs the following processes: If it is determined that the person is in front of the second camera, a specific head image that meets the conditions is selected from the head images representing the head of the person appearing in the image captured by the first camera. as well as The specific head image is compared with the image captured by the second camera, and the identifier obtained as the authentication result is used to establish a correspondence with the specific person corresponding to the specific head image that has a similarity of more than a threshold.

10. A person authentication device, which is connected to a first camera and a second camera, wherein the first camera takes a full-body picture of one or more persons in a monitored area, and the second camera takes a picture of the head of each person when authenticating them at the entrance of the monitored area, wherein... The person authentication device includes: The selection unit, upon recognizing that the person is present in front of the second camera, selects a specific head image that meets certain conditions from head images representing the head of the person appearing in the image captured by the first camera; and A corresponding unit is established, the specific head image is compared with the image captured by the second camera, and the identifier obtained as the authentication result is associated with the specific person corresponding to the specific head image that has a similarity of more than a threshold.

11. The personal authentication device according to claim 10, characterized in that, If a person is identified as being in front of the second camera based on an image captured by the first camera, the selection unit selects a specific head image from the head images representing the head of the person appearing in the image captured by the first camera, wherein the head orientation is similar to that of the head image appearing in the image captured by the second camera.

12. The personal authentication device according to claim 10 or 11, characterized in that, If a person is identified as being in front of the second camera based on an image captured by the first camera, the selection unit detects the head of the person appearing in the image captured by the first camera, saves the information in a database until the information has expired, and deletes the information that has expired from the database. The information is obtained by establishing a correspondence between a head image representing the detected head, a head direction inferred from the head image, and a temporary identifier for identifying the person corresponding to the head image per unit time.

13. The personal authentication device according to claim 10 or 11, characterized in that, If the person in front of the second camera is identified as a single individual based on the image captured by the first camera, the selection unit avoids selecting the specific head image that meets the condition.

14. The personal authentication device according to claim 10 or 11, characterized in that, If, based on an image captured by the first camera, the person is identified as existing at a distance greater than a separation threshold in front of the second camera, the selection unit avoids selecting a specific head image that meets the condition.

15. The personal authentication device according to claim 10 or 11, characterized in that, The selection unit determines whether to select a specific head image that meets the conditions based on the correlation between the facilities in the monitored area and the types of clothing commonly used in the facilities, and avoids selecting the specific head image that meets the conditions if the correlation is higher than a threshold correlation.

16. The personal authentication device according to claim 10 or 11, characterized in that, The second camera is a facial authentication camera that authenticates the face of the person being identified.

17. The personal authentication device according to claim 10 or 11, characterized in that, The second camera is positioned within the shooting range of the first camera.

Images