WO2022176294A1 - 情報処理装置、情報処理方法、照合システム、プログラム及び記録媒体 - Google Patents
情報処理装置、情報処理方法、照合システム、プログラム及び記録媒体 Download PDFInfo
- Publication number
- WO2022176294A1 WO2022176294A1 PCT/JP2021/042538 JP2021042538W WO2022176294A1 WO 2022176294 A1 WO2022176294 A1 WO 2022176294A1 JP 2021042538 W JP2021042538 W JP 2021042538W WO 2022176294 A1 WO2022176294 A1 WO 2022176294A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- registration
- similarity
- registration candidate
- image
- Prior art date
Links
- 230000010365 information processing Effects 0.000 title claims abstract description 119
- 238000012795 verification Methods 0.000 title claims abstract description 18
- 238000003672 processing method Methods 0.000 title claims description 49
- 238000012360 testing method Methods 0.000 claims abstract description 129
- 238000004364 calculation method Methods 0.000 claims abstract description 36
- 238000013500 data storage Methods 0.000 claims abstract description 36
- 238000000034 method Methods 0.000 description 34
- 230000006870 function Effects 0.000 description 28
- 230000008569 process Effects 0.000 description 22
- 238000012545 processing Methods 0.000 description 20
- 238000003384 imaging method Methods 0.000 description 18
- 238000010586 diagram Methods 0.000 description 14
- 230000001815 facial effect Effects 0.000 description 13
- 238000004891 communication Methods 0.000 description 8
- 238000001514 detection method Methods 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 210000003462 vein Anatomy 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000003542 behavioural effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/50—Maintenance of biometric data or enrolment thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/21—Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/2117—User registration
Definitions
- This disclosure relates to an information processing device, an information processing method, a verification system, a program, and a recording medium.
- Patent Document 1 describes a personal authentication system using face data and voice data. Further, Patent Document 2 describes an authentication device that performs authentication using a plurality of matching images. Patent Literature 3 describes a technique for registering appropriate biometric information according to changes over time in biometric information.
- JP-A-2000-148985 JP 2008-059533 A Japanese Patent Application Laid-Open No. 2010-231320
- biometric authentication using image data it is determined whether or not the person in the acquired image is the same as the person registered in the database by comparing the acquired image with the image registered in the database. will be Therefore, an image of a person to be authenticated is registered in advance in the database of the authentication device.
- a registered image has a low image quality, it becomes difficult to distinguish it from another person, which may cause false authentication.
- An object of this disclosure is to provide an information processing device, an information processing method, a verification system, a program, and a recording medium that can mechanically select biometric information data suitable for verification of biometric information from registration candidate data. .
- a similarity calculation unit for calculating a similarity of each of a plurality of registration candidate data each including biometric information of the same person to test data including the biometric information of the person; registration selecting, from among the plurality of registration candidate data, the registration candidate data having the higher degree of similarity to the test data from among the plurality of registration candidate data as registration data to be registered in a data storage unit for matching the person.
- An information processing apparatus having a data selection unit is provided.
- a data acquisition device that acquires biometric information data of a person, a storage device in which a plurality of biometric information data of a plurality of persons are registered, and the biometric data acquired by the data acquisition device
- a similarity calculation unit that calculates a similarity between the information data and the plurality of biometric information data registered in the storage device, and a person represented by the biometric information data acquired by the data acquisition device is registered in the storage device.
- an information processing apparatus having a matching unit that determines whether or not the person is the same person based on the similarity, wherein the similarity calculation unit calculates the biometric information of the same person.
- the information processing device is further configured to calculate a degree of similarity to test data including biometric information of the person for each of a plurality of registration candidate data included in each, and the information processing device stores the The registration data to be registered in the apparatus are selected based on the degree of similarity of each of the plurality of registration candidate data to the test data and the degree of similarity to each of the plurality of biometric information data of the plurality of persons.
- a verification system further comprising a registration data selection unit that selects from the registration candidate data.
- a computer calculates a degree of similarity for each of a plurality of registration candidate data each including the same person's biometric information to test data including the person's biometric information; Among the plurality of registration candidate data, registration candidate data having a high degree of similarity to the test data is selected from the plurality of registration candidate data as registration data to be registered in a data storage unit for matching the person.
- a program is provided that functions as a means.
- FIG. 1 is a block diagram showing a schematic configuration of an information processing apparatus according to a first embodiment
- FIG. 4 is a flow chart showing an information processing method according to the first embodiment
- 1 is a block diagram showing a hardware configuration example of an information processing apparatus according to a first embodiment
- FIG. 8 is a flow chart showing an information processing method according to the second embodiment
- 9 is a flow chart showing an information processing method according to a third embodiment
- FIG. 10 is a flow chart showing an information processing method according to a fourth embodiment; It is a figure which shows the application example of the information processing method by 4th Embodiment.
- FIG. 11 is a block diagram showing a schematic configuration of a collation system according to a fifth embodiment;
- FIG. 12 is a block diagram showing a schematic configuration of an information processing apparatus according to a sixth embodiment;
- FIG. 21 is a block diagram showing a schematic configuration of a collation system according to a seventh embodiment;
- FIG. 1 is a block diagram showing a schematic configuration of an information processing apparatus according to this embodiment.
- FIG. 2 is a flowchart showing an information processing method according to this embodiment.
- FIG. 3 is a diagram showing an application example of the information processing method according to this embodiment.
- FIG. 4 is a block diagram showing a hardware configuration example of the information processing apparatus according to this embodiment.
- the information processing apparatus 100 has a function of selecting an image suitable for identification with another person from a plurality of images (registration candidate images) showing the same person and registering it in a database.
- the information processing apparatus 100 includes an image acquisition unit 110, a matching score calculation unit 120, a registered image selection unit 130, an image registration unit 140, and a data storage unit 150, as shown in FIG. obtain.
- the image acquisition unit 110 is a functional block having a function of acquiring image data representing biological information from an external device (not shown), such as an image acquisition device.
- Image data representing biometric information can be, for example, face image data, fingerprint image data, vein image data, iris image data, and the like.
- the image acquisition device may be an imaging device having a function of acquiring a still image or a moving image.
- the matching score calculation unit 120 is a functional block having a function of matching each of the plurality of registration candidate images acquired by the image acquisition unit 110 with the test image and calculating a matching score representing the degree of similarity to the test image.
- a registration candidate image is an image acquired by the image acquisition unit 110 .
- a test image is an image that is commonly used as a matching image when calculating a matching score for each registration candidate image.
- the registered image selection unit 130 is a functional block having a function of selecting an image suitable for identification with another person from among a plurality of registration candidate images based on the matching score calculated by the matching score calculation unit 120 .
- the image registration unit 140 is a functional block having a function of registering the registration candidate image selected by the registration image selection unit 130 in the data storage unit 150 as registration data.
- the data storage unit 150 is a functional block having a function of holding registration candidate images (registration data) selected by the registration image selection unit 130 .
- the data storage unit 150 may be a part of the information processing device 100 as shown in FIG. 1, or may be another device (storage device) provided outside the information processing device 100. .
- the image acquisition unit 110 acquires a plurality of image data including a face image of a person to be registered from an external device.
- one image is set as a test image, and the remaining plurality of images are set as registration candidate images (step S101).
- a plurality of registration candidate images may be, for example, images captured by an imaging device that constitutes part of the verification system.
- the test image may be one of the images obtained in the same manner as the plurality of registration candidate images, or may be a separately prepared face image of a person to be registered.
- the plurality of registration candidate images may include images captured under different imaging conditions, such as lighting brightness.
- test image When selecting a test image from a plurality of registration candidate images, it is preferable to fix the test image to a specific one from the viewpoint of fair comparison between the registration candidate images. However, if the effect of changing the test image can be considered small, the test image does not necessarily have to be fixed to one image, and the test image was selected from the remaining registered candidate images excluding the registered candidate image under evaluation. Any registration candidate image may be used as the test image.
- a plurality of registration candidate images are selected from the following viewpoints. is desirable.
- the image of the person you are trying to register For example, an image for which there is a possibility that a different person has taken the place of the person in the middle of shooting, and the person cannot be sure that the person is the actual person is excluded from the registration candidate images.
- face recognition whether or not there is a possibility that a person has been impersonated by another person is determined by the fact that a person disappears from the field of view of the camera or two or more persons are captured during the period from the start of the registration operation to the acquisition of candidate images. , etc. can be used as criteria for determination.
- the former can be realized using video tracking technology, and the latter using a face detection mechanism.
- iris authentication it can be realized by a similar method. If there are a plurality of cameras, it is sufficient if at least one camera confirms that the person has not been replaced.
- the feature to be registered is reflected in the image.
- face recognition the face must be captured
- iris recognition the iris must be captured.
- biometric authentication devices have a face detection function for face authentication and an iris detection function for iris authentication, so they can be used as they are.
- the image quality is good.
- An example of good image quality is being in focus. Whether or not the image is in focus can be confirmed by a known method (for example, performing a Fourier transform of the image to check the presence or absence of high frequency components).
- the matching score calculation unit 120 sequentially calculates a matching score St for each of the plurality of registration candidate images acquired by the image acquisition unit 110 (step S102).
- the method of calculating the matching score St is not particularly limited.
- facial feature amounts which are parameters representing facial features, are extracted from facial images included in registered candidate images.
- a facial feature amount is a vector amount, and is a combination of scalar amount components that express features of a facial image.
- the component of the feature quantity is not particularly limited, and various types can be used.
- positional relationships such as distances and angles between feature points set at the center or end points of facial organs such as eyes, nose, and mouth, curvature of the contour of the face, color distribution of the face surface, and so on.
- a grayscale value or the like can be used.
- the number of components of the feature amount is also not particularly limited, and can be set as appropriate according to the required collation accuracy, processing speed, and the like.
- the facial feature amount extracted from the facial image included in the registration candidate image is compared with the facial feature amount extracted from the facial image included in the test image, and a matching score St representing the similarity of these facial feature amounts is calculated.
- the matching score St is a numerical value between 0 and 1, and is closer to 1 as the similarity of the facial feature amount is higher, and is closer to 0 as the similarity is lower.
- step S103 the registered image selection unit 130 initializes parameters used for image selection (step S103).
- 0 is registered in the variable representing the maximum score Sm
- null data is registered in the variable representing the information of the candidate data, for example, the character string variable storing the file name of the registration candidate image.
- step S103 may be performed before image selection processing (steps S104 to S106), which will be described later, and may be performed immediately before or after step S101, for example.
- the registration image selection unit 130 sequentially selects a plurality of registration candidate images, and repeats the processing from step S104 to step S106.
- step S104 it is determined whether or not the matching score St calculated in step S102 satisfies a predetermined condition for the selected registration candidate image.
- the predetermined condition here is that the collation score St is greater than the minimum value (threshold value) suitable for collation.
- the process proceeds to step S105. If the result of determination in step S104 is that the collation score St does not satisfy the predetermined condition (No), the process proceeds to the next registration candidate image processing (step S104).
- the threshold is 0.50, in the example of FIG. 3, it is determined that the registration candidate image with the file name "A004" does not satisfy the predetermined condition, and the other registration candidate images satisfy the predetermined condition. will be judged.
- Executing step S104 is particularly effective in eliminating selection of an inappropriate image as a registered image when the population of registered candidate images is small.
- the test image and the registration candidate image are images of the same person, and it is generally assumed that the matching score St is a value equal to or greater than the threshold.
- the determination in step S104 does not necessarily have to be performed.
- step S105 it is determined whether or not the matching score St calculated in step S102 for the selected registration candidate image is greater than the maximum score Sm registered at that time. If the result of determination in step S105 is that the collation score St is greater than the maximum score Sm (Yes), the process proceeds to step S106. If the result of determination in step S105 is that the collation score St is equal to or less than the maximum score Sm (No), the process proceeds to the next registration candidate image processing (step S104).
- the maximum score Sm is finally obtained as the maximum of the matching scores St of the plurality of registration candidate images. is registered. Further, as the candidate data, information of the registered candidate image having the maximum collation score St is registered.
- step S104 After the processing from step S104 to step S106 is completed for all registration candidate images in the registration image selection unit 130, the process proceeds to step S107.
- the image registration unit 140 stores the registration candidate image having the maximum matching score St in the database of the data storage unit 150 as registration data.
- the registration data stored in the database of data storage unit 150 may include the image data acquired in step S101.
- the registration data stored in the database of the data storage unit 150 may include the feature amount data calculated in step S102.
- one registered candidate image with the highest matching score for the test image is selected as registered data, but multiple registered candidate images with higher matching scores may be selected as registered data.
- the quality of the registration candidate image is determined mechanically based on the matching score for the test image. Therefore, the determination of selecting an appropriate registered image from among a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether the registered candidate images are good or bad. Thereby, a database suitable for biometric authentication can be constructed.
- the information processing method of the present embodiment can be preferably applied, for example, when it is assumed that the match score of a registered candidate image with respect to a test image is sufficiently higher than the match score of a registered image with another person's image.
- the match score of the registered image is higher than the matching score of the image of another person, for example, the camera that acquires the image is high performance, the acquired test image and the registered candidate image are high quality, and the image of the person
- a sufficiently high value can be expected as a matching score between the two.
- the information processing apparatus 100 according to this embodiment can be realized by a hardware configuration similar to that of a general information processing apparatus. That is, the information processing apparatus 100 can be configured including a processor 200, a main storage unit 202, a communication unit 204, and an input/output interface unit 206, as shown in FIG. 4, for example.
- the processor 200 is a control/arithmetic device that controls each functional block of the information processing device 100 and performs arithmetic processing.
- the processor 200 uses one of CPU (Central Processing Unit), GPU (Graphics Processing Unit), FPGA (Field-Programmable Gate Array), DSP (Digital Signal Processor), and ASIC (Application Specific Integrated Circuit). may be used, or a plurality of them may be used in parallel.
- CPU Central Processing Unit
- GPU Graphics Processing Unit
- FPGA Field-Programmable Gate Array
- DSP Digital Signal Processor
- ASIC Application Specific Integrated Circuit
- the main storage unit 202 is a storage unit used as a data work area and a data temporary save area, and is composed of a memory such as a RAM (Random Access Memory).
- a communication unit 204 is an interface for transmitting and receiving data via a network.
- the input/output interface unit 206 is an interface for transmitting and receiving data by connecting to an external output device 210, input device 212, storage device 214, and the like.
- the processor 200 , main storage unit 202 , communication unit 204 and input/output interface unit 206 are interconnected by a system bus 208 .
- the main storage unit 202 can be used as a work area for executing calculations when calculating the matching score St and extracting registered data.
- the processor 200 functions as a control unit that controls these arithmetic processes, and functions as an image acquisition unit 110, a matching score calculation unit 120, a registered image selection unit 130, and an image registration unit 140 together with a main storage unit 202 and an input/output interface unit 206.
- play the role of Storage device 214 can be used as data storage section 150 that stores registration data selected by registration image selection section 130 .
- the communication unit 204 is a communication interface based on standards such as Ethernet (registered trademark) and Wi-Fi (registered trademark), and is a module for communicating with other devices.
- Registration data stored in storage device 214 may be configured to be received from another device via communication unit 204 .
- registration data built in a device other than the information processing device of this embodiment, such as an authentication device can be received via the communication unit 204 and stored in the storage device 214 .
- the registration data stored in this way can be used for matching processing, and can also be used for extracting candidate data in the information processing method according to the second embodiment described later.
- the input device 212 is a keyboard, mouse, touch panel, etc., and is used by the user to input predetermined information to the information processing device 100 .
- the input device 212 can also be used as means for inputting registration candidate images and test images. For example, if the registration candidate image or test image is a two-dimensional image, an image reading device can be used as the input device 212 .
- the output device 210 is a display device, a printer device, or the like, and is used to notify the user of the progress of computation and the result of computation.
- the storage device 214 can be configured, for example, by a hard disk device or the like composed of a non-volatile memory such as a ROM (Read Only Memory), a magnetic disk, or a semiconductor memory.
- each part of the information processing apparatus 100 can be realized in terms of hardware by implementing circuit components, which are hardware components such as LSI (Large Scale Integration) incorporating programs.
- a program that provides the function can be stored in the storage device 214, loaded into the main storage unit 202, and executed by the processor 200, thereby realizing software.
- Image analysis techniques such as Fourier transform, may be used to numerically evaluate whether or not an image is in focus, but it is difficult to indicate what the numerical value should be. Focus is not the only factor that determines whether a registered image is good or bad. In other words, just by observing the behavior of the algorithm used for image authentication determination from the outside, it is possible to know all the characteristics that an image suitable as a registered image has and to what extent those characteristics should be satisfied. is generally difficult.
- FIG. 5 is a flowchart showing an information processing method according to this embodiment.
- FIG. 6 is a diagram showing an application example of the information processing method according to this embodiment.
- the image acquisition unit 110 acquires a plurality of image data including a face image of a person to be registered from an external device.
- one image is set as a test image, and the remaining plurality of images are set as registration candidate images (step S201).
- Registration candidate images and test images are the same as in the first embodiment.
- the matching score calculation unit 120 sequentially selects a plurality of registration candidate images, and repeats the processes of steps S202 and S203. Specifically, in step S202, a match score St for the selected registration candidate image with respect to the test image is calculated. Also, in step S203, for the selected registration candidate image, a matching score Sr is calculated for each piece of registered data of another person registered in the database of the data storage unit 150. FIG. The method for calculating the matching score Sr is the same as the method for calculating the matching score St.
- the registration data of another person is not registered in the database of the data storage unit 150 or when the population of the registration data of the other person is small, the data of the other person acquired in a different environment is used instead of the registration data. good too.
- Examples of such data include, but are not limited to, data disclosed on public databases, data obtained at other locations, and the like.
- step S204 the registered image selection unit 130 initializes parameters used for image selection (step S204).
- 0 is registered in the variable Sm representing the maximum score
- null data is registered in the variable representing the information of the candidate data, for example, the character string variable storing the file name of the registration candidate image.
- step S204 may be performed before image selection processing (steps S205 to S208), which will be described later, and may be performed immediately before or after step S201, for example.
- the registration image selection unit 130 sequentially selects a plurality of registration candidate images, and repeats the processing from step S205 to step S208.
- step S205 it is determined whether or not the matching score St calculated in step S202 satisfies a predetermined condition for the selected registration candidate image.
- the determination in step S202 is the same as the determination in step S104 in the first embodiment.
- the process proceeds to step S206. If the result of determination in step S205 is that the collation score St does not satisfy the predetermined condition (No), the process proceeds to the next registration candidate image processing (step S205).
- the threshold is 0.50, in the example of FIG. 3, it is determined that the registration candidate image with the file name "A004" does not satisfy the predetermined condition, and the other registration candidate images satisfy the predetermined condition. will be judged.
- step S206 it is determined whether or not the matching score St calculated in step S202 is greater than the matching score Sr calculated in step S203 for the selected registration candidate image. If the result of determination in step S206 is that the matching score St is greater than any matching score Sr calculated in step S203 (Yes), the process proceeds to step S207. If the result of determination in step S206 is that the matching score St is equal to or less than at least one matching score Sr (No), the process proceeds to the next registration candidate image (step S205).
- step S205 it is determined that the registration candidate image with the file name "A004" does not satisfy the condition of step S206.
- step S207 it is determined whether or not the matching score St calculated in step S202 for the selected registration candidate image is greater than the maximum score Sm registered at that time. If the result of determination in step S207 is that the collation score St is greater than the maximum score Sm (Yes), the process proceeds to step S208. If the result of determination in step S207 is that the collation score St is less than or equal to the maximum score Sm (No), the process proceeds to the next registration candidate image (step S205).
- step S208 the matching score St of the registration candidate image determined to be greater than the maximum score Sm in step S207 is overwritten as the maximum score Sm. Also, the information of the registration candidate image is registered as candidate data.
- the maximum score Sm is finally obtained as the maximum matching score St of the plurality of registration candidate images.
- a collation score St is registered.
- the candidate data information of the registered candidate image having the maximum collation score St is registered.
- the matching score St of the registration candidate image with the file name "A005" is the maximum score Sm, and the information of this registration candidate image is registered as candidate data.
- the image data selected in this manner is the image data with the highest matching score St among the registration candidate images, and has the higher matching score St than any registered data registered in the database of the data storage unit 150. become a thing.
- step S208 After the processing from step S205 to step S208 is completed for all registration candidate images in the registration image selection unit 130, the process proceeds to step S209.
- step S209 the image registration unit 140 stores the registration candidate image having the maximum matching score St in the database of the data storage unit 150 as registration data.
- one registration candidate image with the highest matching score for the test image and registered images of others is selected as registration data, but multiple registration candidate images with high matching scores are selected as registration data.
- step S209 If the candidate data remains empty in step S209, it is possible that the person appearing in the registration candidate image has already been registered in the database of the data storage unit 150. In this case, a notification such as "already registered” may be displayed on the screen of the display device.
- the quality of the registration candidate image is determined mechanically based on the matching score for the test image. Therefore, the determination of selecting an appropriate registered image from among a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether the registered candidate images are good or bad.
- the matching score St for the test image the matching score Sr for other person's registered data is also used as a judgment criterion, so that an image having a discriminative power with respect to other person's registered data is efficiently extracted from registered candidate images. can do. Thereby, a database suitable for biometric authentication can be constructed.
- the registration data of another person read from the database of the data storage unit 150 may be registration data used in actual verification processing and authentication processing.
- the determination of selecting an appropriate registered image from a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether a registered candidate image is good or bad. can be done.
- the matching score for the test image the matching score for other person's registered data is also used as a judgment criterion. Therefore, it is possible to efficiently extract an image having discriminative power from other person's registered data from registered candidate images. can be done. Thereby, a database suitable for biometric authentication can be constructed.
- FIG. 7 is a flow chart showing an information processing method according to this embodiment.
- FIG. 8 is a diagram showing an application example of the information processing method according to this embodiment.
- the matching score St for each of a plurality of registered candidate images is calculated for one test image, and the registered candidate image having the maximum matching score St is used as the registered image.
- the image extracted as the registered image will also be another image.
- the image selected as the test image is unclear or otherwise inappropriate as a criterion for calculating the matching score, it may not be possible to extract an image that is suitable as registration data from among the candidate images for registration. is assumed.
- the image acquisition unit 110 acquires a plurality of image data including a face image of a person to be registered from an external device.
- a plurality of image data acquired in this manner are used as registration candidate images (step S301).
- Each of the plurality of acquired image data is used as a registration candidate image and also as a test image.
- the registration candidate images By using the registration candidate images as test images, the number of images to be acquired can be reduced. If there is no particular limit on the number of images to be acquired, a plurality of test images may be prepared separately from the registration candidate images.
- the matching score calculation unit 120 sequentially selects a plurality of registration candidate images, and repeats the processes of steps S302 and S303. Specifically, in step S302, the selected registration candidate image is designated as the test image. Next, in step S303, for each of the other registration candidate images, a matching score St for the designated test image is calculated.
- step S304 for each registered candidate image designated as a test image, the matching scores St calculated for each of the other registered candidate images are arranged in descending order of score, and score ranking is given.
- FIG. 8 shows the comparison results calculated by the procedures in steps S302 and S303 for six registration candidate images with file names "A001", “A002", “A003", “A004", “A005", and “A006”.
- the score St Numerical values in parentheses indicate rankings (score rankings) when matching scores St calculated for the same test image are arranged in descending order. For example, when the registered candidate image with the file name “A002” is selected as the test image, the matching score of the registered candidate image with the file name “A001” is 0.87, and the score order is 4. Also, the collation score of the registration candidate image with the file name “A003” is 0.92, and the score order is 3.
- the matching score of the registered candidate image with the file name “A004” is 0.74, and the score order is 5.
- the matching score of the registered candidate image with the file name “A005” is 0.97, and the score order is 1.
- the matching score of the registration candidate image with the file name “A006” is 0.93, and the score order is 2.
- the registration image selection unit 130 calculates, for each registration candidate image, the sum of the score ranks of other registration candidate images obtained as test images (step S305).
- the sum of the score ranks of the registration candidate image with the file name “A001” is 20, and the sum of the score ranks of the registration candidate image with the file name “A002” is 9.
- the sum of the score ranks of the registration candidate image with the file name “A003” is 13, and the sum of the score ranks of the registration candidate image with the file name “A004” is 25.
- the sum of the score ranks of the registration candidate image with the file name “A005” is 7, and the sum of the score ranks of the registration candidate image with the file name “A006” is 16.
- the registration image selection unit 130 selects, as registration data, the registration candidate image with the smallest sum of the score ranks calculated in step S305 from among the plurality of registration candidate images (step S306).
- the registration candidate image with the file name "A005" is selected as the registration data.
- selecting the registered candidate image with the smallest sum of score ranks means selecting the registered candidate image with the highest similarity to other registered candidate images. This makes it possible to extract a more appropriate registered image than when extracting a registered image based on one test image.
- the image registration unit 140 stores the registration candidate image selected in step S305 as registration data in the data storage unit 150 (step S307).
- the quality of the registration candidate image is determined mechanically based on the matching score for the test image. Therefore, the determination of selecting an appropriate registered image from among a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether the registered candidate images are good or bad. Also, using a plurality of test images, a matching score is calculated for each of the registration candidate images. Therefore, compared to the case where one image is used as a test image, it is possible to extract a more optimal registered image.
- all registered candidate images are used as test images, but a predetermined number of registered candidate images, ie, two or more, out of a plurality of registered candidate images may be used as test images. If the number of matching scores calculated for each registration candidate image is different, the average value of the score ranks may be used instead of the sum of the score ranks.
- step S104 in the first embodiment and the determination in steps S205 and S206 in the second embodiment are not performed, but these determination steps may be performed.
- the determination of selecting an appropriate registered image from a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether a registered candidate image is good or bad. can be done. Also, using a plurality of test images, a matching score is calculated for each of the registration candidate images. Therefore, compared to the case where one image is used as a test image, it is possible to extract a more optimal registered image. Thereby, a database suitable for biometric authentication can be constructed. Moreover, since the test image is specified from among a plurality of registration candidate images, the number of images to be acquired can be reduced.
- FIG. 9 is a flow chart showing an information processing method according to this embodiment.
- FIG. 10 is a diagram showing an application example of the information processing method according to this embodiment.
- the image acquisition unit 110 acquires a plurality of face images of a person to be registered from an external image acquisition device or the like.
- a plurality of face images acquired in this manner are used as registration candidate images (step S401).
- Each of the plurality of acquired image data is used as a registration candidate image and also as a test image.
- the registration candidate images By using the registration candidate images as test images, the number of images to be obtained can be reduced. If there is no particular limit on the number of images to be acquired, a plurality of test images may be prepared separately from the registration candidate images.
- the matching score calculation unit 120 sequentially selects a plurality of registration candidate images, and repeats the processes of steps S402 and S403. Specifically, in step S402, the selected registration candidate image is designated as the test image. Next, in step S403, for each of the other registration candidate images, a matching score St for the designated test image is calculated.
- FIG. 10 shows the comparison results calculated by the procedures in steps S302 and 303 for six registration candidate images with file names "A001", “A002", “A003", “A004", “A005", and "A006".
- This is an example of summarizing the scores St.
- the matching score St of the registration candidate image with the file name "A001” is 0.87.
- the matching score St of the registration candidate image with the file name "A003” is 0.92.
- the matching score St of the registration candidate image with the file name "A004" is 0.74.
- the matching score St of the registration candidate image with the file name "A005" is 0.97.
- the matching score St of the registration candidate image with the file name "A006" is 0.93.
- the registration image selection unit 130 calculates the sum of matching scores St obtained by using other registration candidate images as test images (step S404).
- the sum of the matching scores St for the registered candidate image with the file name "A001" is 4.09
- the sum of the matching scores St for the registered candidate image with the file name "A002" is 4.30.
- the sum of matching scores St for the registered candidate image with the file name "A003" is 4.27
- the sum of the matching scores St for the registered candidate image with the file name "A004" is 3.53.
- the sum of matching scores St for the registered candidate image with the file name "A005" is 4.33
- the sum of the matching scores St for the registered candidate image with the file name "A006" is 4.19.
- the registration image selection unit 130 selects, as registration data, the registration candidate image having the maximum sum of the matching scores St calculated in step S404 from among the plurality of registration candidate images (step S405).
- the registration candidate image with the file name "A005" is selected as the registration data.
- the matching score St increases as the similarity to the test image increases. Therefore, selecting the registered candidate image with the largest sum of matching scores St means selecting the registered candidate image with the highest similarity to other registered candidate images. This makes it possible to extract a more appropriate registered image than when extracting a registered image based on one test image.
- the image registration unit 140 stores the registration candidate image selected in step S404 as registration data in the data storage unit 150 (step S405).
- the quality of the registration candidate image is determined mechanically based on the matching score for the test image. Therefore, the determination of selecting an appropriate registered image from among a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether the registered candidate images are good or bad. Also, using a plurality of test images, a matching score is calculated for each of the registration candidate images. Therefore, compared to the case where one image is used as a test image, it is possible to extract a more optimal registered image.
- all registered candidate images are used as test images, but a predetermined number of registered candidate images, ie, two or more, out of a plurality of registered candidate images may be used as test images. If the number of matching scores calculated for each registration candidate image is different, the average value of matching scores may be used instead of the sum of matching scores.
- step S104 in the first embodiment and the determination in steps S205 and S206 in the second embodiment are not performed, but these determination steps may be performed.
- the determination of selecting an appropriate registered image from a plurality of registered candidate images can be made faster and more accurately than when a human visually determines whether a registered candidate image is good or bad. can be done. Also, using a plurality of test images, a matching score is calculated for each of the registration candidate images. Therefore, compared to the case where one image is used as a test image, it is possible to extract a more optimal registered image. Thereby, a database suitable for biometric authentication can be constructed. Moreover, since the test image is specified from among a plurality of registration candidate images, the number of images to be obtained can be reduced.
- FIG. 11 is a block diagram showing a configuration example of a matching system according to this embodiment.
- a matching system 1000 can be configured by an information processing device 100, an imaging device 300, and a storage device 400, as shown in FIG.
- the imaging device 300 is connected to the information processing device 100 .
- the imaging device 300 may be a part of the information processing device 100, or may be connected to the information processing device 100 via a network or the like.
- the storage device 400 is connected to the information processing device 100 .
- the storage device 400 may be part of the information processing device 100 (for example, the data storage unit 150 in the first embodiment), or may be connected to the information processing device 100 via a network or the like.
- the information processing apparatus 100 further has a matching unit 160 in addition to the image acquiring unit 110, matching score calculating unit 120, registered image selecting unit 130, and image registering unit 140 described in the first embodiment.
- the collation unit 160 collates the image data of the person captured by the imaging device 300 with the image data of the person registered in the storage device 400 to determine whether the person captured by the imaging device 300 is the person registered in the storage device 400 . It is a functional block having a function of determining whether or not.
- the determination result by the matching unit 160 can be notified to the outside via the output device 210 . Also, the determination result by the collating unit 160 can be used for authentication in a gate device or the like.
- the matching system 1000 has a function of selecting and registering image data of a person photographed by the imaging device 300, and a function of determining whether or not the person photographed by the imaging device 300 is a registered person. Prepare.
- the function of selecting and registering the image data of a person captured by the imaging device 300 is as described in the first to fourth embodiments, so the description is omitted here.
- the function of determining whether or not a person photographed by the imaging device 300 is a registered person is performed by an image acquiring unit 110, a matching score calculating unit 120, and a matching unit 160 among the constituent elements of the information processing device 100. It can be implemented by
- the imaging device 300 photographs a person in the field of view and outputs image data to the information processing device 100 .
- the image acquisition unit 110 of the information processing device 100 acquires image data output from the imaging device 300 and outputs the image data to the matching score calculation unit 120 .
- the match score calculator 120 calculates a match score Sc for the image data of the person registered in the database of the storage device 400 for the image data acquired from the image acquirer 110 .
- the matching score calculator 120 used to calculate the matching score St of the registration candidate image with respect to the test image can be used to calculate the matching score Sc.
- the algorithm for calculating the matching score Sc can select the image data that can best distinguish between the person's image and the other person's image from the registered candidate images, and the matching accuracy can be improved. can.
- the matching unit 160 extracts image data with the maximum matching score Sc calculated by the matching score calculation unit 120 from the image data of the registered person. Then, when the matching score Sc of the extracted image data is larger than a predetermined threshold, it is determined that the person photographed by the imaging device 300 is the same person as one of the persons registered in the database. On the other hand, when the collation score Sc of the extracted image data is equal to or less than the predetermined threshold value, it is determined that the person photographed by the imaging device 300 is not registered in the database.
- the matching score calculation unit 120 and registered data used in the actual matching process are used as they are, and the image with the highest matching score among the registration candidate images is selected as the registered data. Therefore, in order to construct a matching system having a function of selecting registered images, there is no need to separately prepare means for estimating the image expected to give the highest matching score from registered candidate images. It is possible to simplify the system and select registered images suitable for the authentication system.
- the matching system according to this embodiment can be applied for various purposes.
- the verification system according to this embodiment is not particularly limited, but can be used, for example, to identify a purchaser in settlement at a store.
- an image expected to give the highest matching score is estimated from registered candidate images. There is no need to prepare a separate means, the system can be simplified, and it is possible to select registered images suitable for the authentication system.
- FIG. 12 is a block diagram showing a schematic configuration of the information processing apparatus according to this embodiment.
- the information processing apparatus 500 has at least a similarity calculation section 510 and a registration data selection section 530, as shown in FIG.
- the similarity calculation unit 510 has a function of calculating the similarity of each of a plurality of registration candidate data each containing the same person's biometric information to the test data containing the person's biometric information.
- the matching score calculator 120 described in the first to fifth embodiments is an example of the similarity calculator 510 .
- Registered data selection unit 530 selects registration candidate data having a high degree of similarity to the test data among the plurality of registration candidate data as registration data to be registered in the data storage unit for verification of the person. It has a function to choose from.
- the registration image selection section 130 described in the first to fifth embodiments is an example of the registration data selection section 530 .
- the determination of selecting appropriate registration data from a plurality of registration candidate data can be made faster and more accurately than when a human visually determines whether a registration candidate image is good or bad. can be done.
- a database suitable for biometric authentication can be constructed.
- FIG. 13 is a block diagram showing a schematic configuration of a matching system according to this embodiment.
- a collation system 1000 includes at least an information processing device 500, a data acquisition device 600, and a storage device 700, as shown in FIG.
- the information processing device 500 has at least a similarity calculation section 510 , a matching section 520 , and a registered data selection section 530 .
- the data acquisition device 600 has a function of acquiring biometric information data of a person.
- the imaging device 300 described in the fifth embodiment is an example of the data acquisition device 600 .
- the storage device 700 has a function of registering a plurality of biometric information data of a plurality of persons.
- the data storage unit 150 described in the first to fourth embodiments and the storage device 400 described in the fifth embodiment are examples of the storage device 700 .
- the information processing device 500 includes a similarity calculation unit 510 that calculates the similarity between the biometric information data acquired by the data acquisition device 600 and a plurality of biometric information data registered in the storage device 700, and and a matching unit 520 that determines whether or not the person represented by the obtained biometric information data is the person registered in the storage device 700 based on the degree of similarity.
- the matching score calculator 120 described in the first to fifth embodiments is an example of the similarity calculator 510 .
- the collation section 160 described in the fifth embodiment is an example of the collation section 520 .
- the similarity calculation unit 510 is further configured to calculate the similarity of each of a plurality of registration candidate data each containing the same person's biometric information to the test data containing the person's biometric information.
- the information processing apparatus 500 sets the registration data to be registered in the storage device 700 for verification of the person as the similarity of each of the plurality of registration candidate data to the test data, and the plurality of biometric information data of the plurality of persons. and a registration data selection unit 530 that selects from among a plurality of registration candidate data based on the degree of similarity to each of .
- the registration image selection section 130 described in the first to fifth embodiments is an example of the registration data selection section 530 .
- image data is shown as an example of biometric information data to be matched, but biometric information data is not limited to image data.
- the biometric information data may be image data, voiceprint data recording a person's voice, or data representing behavioral characteristics such as walking manner.
- an example of a face image is shown as image data, but the image data is not limited to a face image.
- the image data may be an iris image, a fingerprint image, a palmprint image, a vein image, an image representing the shape of the auricle, or the like.
- the matching scores St and Sr are defined as numerical values between 0 and 1, but the definitions of the matching scores St and Sr are not limited to this. Further, in the above embodiment, the matching scores St and Sr are used as indices representing the degree of similarity, but the matching scores St and Sr may be used as indices representing the degree of dissimilarity. In this case, the greater the degree of difference, the greater the values of the matching scores St and Sr.
- registration candidate images are used as registration candidate images, but the number of registration candidate images is not particularly limited.
- a program for operating the configuration of the embodiment is recorded on a recording medium so as to realize the functions of the above embodiment
- the program recorded on the recording medium is read as code, and executed by a computer. It is included in the scope of the embodiment. That is, a computer-readable recording medium is also included in the scope of each embodiment.
- the program itself is also included in each embodiment.
- a floppy (registered trademark) disk, hard disk, optical disk, magneto-optical disk, CD-ROM, magnetic tape, nonvolatile memory card, and ROM can be used as the recording medium.
- the program recorded on the recording medium alone executes the process, but also the one that operates on the OS and executes the process in cooperation with other software and functions of the expansion board. included in the category of
- (Appendix 1) a similarity calculation unit that calculates the similarity of each of a plurality of registration candidate data each containing the same person's biometric information to test data containing the person's biometric information; Among the plurality of registration candidate data, the registration candidate data having the higher degree of similarity to the test data is selected from the plurality of registration candidate data as registration data to be registered in a data storage unit for matching the person. and a registration data selection unit for performing registration.
- the data storage unit has a plurality of data including biometric information of a plurality of persons different from the person,
- the registration data selection unit selects the registration data based on the degree of similarity to the test data and the degree of similarity of each of the plurality of registration candidate data to each of the plurality of data.
- the information processing device according to appendix 1.
- the registration data selection unit selects, from among the plurality of registration candidate data, registration candidate data having a higher similarity to the test data and higher than the similarity to the plurality of data as the registration data.
- the information processing apparatus according to appendix 2, characterized by:
- Appendix 4 The information processing apparatus according to any one of appendices 1 to 3, wherein the registration data is registration candidate data having the highest degree of similarity to the test data among the plurality of registration candidate data.
- the similarity is a match score for the test data,
- the similarity is a matching score for the test data
- the information processing apparatus according to Supplementary note 7, wherein the registration candidate data having the highest similarity to the other registration candidate data is the registration candidate data having the largest sum of the matching scores with respect to the other registration candidate data.
- Appendix 10 a data acquisition device for acquiring biometric information data of a person; a storage device in which a plurality of pieces of biometric information data of a plurality of persons are registered; a similarity calculation unit that calculates a degree of similarity between the biometric information data acquired by the data acquisition device and the plurality of biometric information data registered in the storage device; and the biometric information data acquired by the data acquisition device.
- the similarity calculation unit is further configured to calculate a similarity of each of a plurality of registration candidate data each including the same person's biometric information to test data including the person's biometric information
- the information processing device stores registration data to be registered in the storage device for verification of the person as the degree of similarity of each of the plurality of registration candidate data to the test data, and the degree of similarity of each of the plurality of persons.
- a verification system further comprising: a registration data selection unit that selects from among the plurality of registration candidate data based on a degree of similarity to each piece of biometric information data.
- Appendix 12 The degree of similarity to the test data, and the degree of similarity of each of the plurality of registration candidate data to each of a plurality of data including biometric information of a plurality of persons different from the person registered in the data storage unit. 12. The information processing method according to appendix 11, wherein the registration data is selected based on .
- Appendix 14 14. The information processing method according to any one of appendices 11 to 13, wherein the registration data is registration candidate data having the highest degree of similarity to the test data among the plurality of registration candidate data.
- Appendix 16 15. The information processing according to any one of appendices 11 to 14, wherein for each of the plurality of registration candidate data, the similarity is calculated using each of all other registration candidate data as the test data. Method.
- the similarity is a match score for the test data, 18.
- the similarity is a match score for the test data, 18.
- Appendix 21 A computer-readable recording medium recording the program according to appendix 20.
- DESCRIPTION OF SYMBOLS 100 Information processing apparatus 110... Image acquisition part 120... Verification score calculation part 130... Registration image selection part 140... Image registration part 150... Data storage part 160... Verification part 200... Processor 202... Main storage part 204... Communication part 206... Input/output interface unit 208 System bus 210 Output device 212 Input device 214 Storage device 300 Imaging devices 400, 700 Storage device 500 Information processing device 510 Similarity calculation unit 520 Verification unit 530 Registration data selection Part 600...Data acquisition device 1000...Authentication system
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Security & Cryptography (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Collating Specific Patterns (AREA)
Abstract
Description
第1実施形態による情報処理装置及び情報処理方法について、図1乃至図4を用いて説明する。図1は、本実施形態による情報処理装置の概略構成を示すブロック図である。図2は、本実施形態による情報処理方法を示すフローチャートである。図3は、本実施形態による情報処理方法の適用例を示す図である。図4は、本実施形態による情報処理装置のハードウェア構成例を示すブロック図である。
本実施形態による情報処理装置100は、一般的な情報処理装置と同様のハードウェア構成によって実現することが可能である。すなわち、情報処理装置100は、例えば図4に示すように、プロセッサ200と、主記憶部202と、通信部204と、入出力インターフェース部206と、を含んで構成され得る。
第2実施形態による情報処理方法について、図5及び図6を用いて説明する。第1実施形態と同様の構成要素には同一の符号を付し、説明を省略し或いは簡潔にする。図5は、本実施形態による情報処理方法を示すフローチャートである。図6は、本実施形態による情報処理方法の適用例を示す図である。
第3実施形態による情報処理方法について、図7及び図8を用いて説明する。第1及び第2実施形態と同様の構成要素には同一の符号を付し、説明を省略し或いは簡潔にする。図7は、本実施形態による情報処理方法を示すフローチャートである。図8は、本実施形態による情報処理方法の適用例を示す図である。
第4実施形態による情報処理方法について、図9及び図10を用いて説明する。第1乃至第3実施形態と同様の構成要素には同一の符号を付し、説明を省略し或いは簡潔にする。図9は、本実施形態による情報処理方法を示すフローチャートである。図10は、本実施形態による情報処理方法の適用例を示す図である。
第5実施形態による照合システムについて、図11を用いて説明する。第1乃至第4実施形態と同様の構成要素には同一の符号を付し、説明を省略し或いは簡潔にする。図11は、本実施形態による照合システムの構成例を示すブロック図である。
第6実施形態による情報処理装置について、図12を用いて説明する。第1乃至第5実施形態と同様の構成要素には同一の符号を付し、説明を省略し或いは簡潔にする。図12は、本実施形態による情報処理装置の概略構成を示すブロック図である。
第7実施形態による情報処理装置について、図13を用いて説明する。第1乃至第6実施形態と同様の構成要素には同一の符号を付し、説明を省略し或いは簡潔にする。図13は、本実施形態による照合システムの概略構成を示すブロック図である。
この開示は、上記実施形態に限らず種々の変形が可能である。
例えば、何れかの実施形態の一部の構成を他の実施形態に追加した例や、他の実施形態の一部の構成と置換した例も、この開示の実施形態である。
同じ人物の生体情報を各々が含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出する類似度計算部と、
前記複数の登録候補データのうち前記テストデータに対する前記類似度が上位の登録候補データを、前記人物の照合のためにデータ記憶部に登録する登録データとして、前記複数の登録候補データの中から選択する登録データ選択部と
を有することを特徴とする情報処理装置。
前記データ記憶部は、前記人物とは異なる複数の人物の生体情報を含む複数のデータを有し、
前記登録データ選択部は、前記テストデータに対する前記類似度と、前記複数の登録候補データの各々の前記複数のデータの各々に対する類似度と、に基づいて前記登録データを選択する
ことを特徴とする付記1記載の情報処理装置。
前記登録データ選択部は、前記複数の登録候補データのうち、前記テストデータに対する類似度が上位であり且つ前記複数のデータに対する前記類似度よりも高い登録候補データを、前記登録データとして選択する
ことを特徴とする付記2記載の情報処理装置。
前記登録データは、前記複数の登録候補データのうち、前記テストデータに対する類似度が最も高い登録候補データである
ことを特徴とする付記1乃至3の何れか1項に記載の情報処理装置。
前記類似度計算部は、前記複数の登録候補データの各々について、他の登録候補データを前記テストデータとして前記類似度を算出する
ことを特徴とする付記1乃至4の何れか1項に記載の情報処理装置。
前記類似度計算部は、前記複数の登録候補データの各々について、他の総ての登録候補データの各々を前記テストデータとして前記類似度を算出する
ことを特徴とする付記1乃至4の何れか1項に記載の情報処理装置。
前記登録データ選択部は、前記複数の登録候補データのうち、他の登録候補データに対する類似度が最も高い登録候補データを、前記登録データとして選択する
ことを特徴とする付記6記載の情報処理装置。
前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアのスコア順位の和が最も小さい登録候補データである
ことを特徴とする付記7記載の情報処理装置。
前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアの和が最も大きい登録候補データである
ことを特徴とする付記7記載の情報処理装置。
人物の生体情報データを取得するデータ取得装置と、
複数の人物の複数の生体情報データが登録された記憶装置と、
前記データ取得装置が取得した前記生体情報データと前記記憶装置に登録されている前記複数の生体情報データとの類似度を算出する類似度計算部と、前記データ取得装置が取得した前記生体情報データが表す人物が前記記憶装置に登録されている人物であるか否かを前記類似度に基づいて判定する照合部と、を有する情報処理装置と、を有する照合システムであって、
前記類似度計算部は、同じ人物の生体情報を各々が含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出するように更に構成されており、
前記情報処理装置は、前記人物の照合のために前記記憶装置に登録する登録データを、前記複数の登録候補データの各々の、前記テストデータに対する前記類似度と、前記複数の人物の前記複数の生体情報データの各々に対する類似度と、に基づいて前記複数の登録候補データの中から選択する登録データ選択部を更に有する
ことを特徴とする照合システム。
同じ人物の生体情報を含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出し、
前記複数の登録候補データのうち前記テストデータに対する類似度が上位の登録候補データを、前記人物の照合のためにデータ記憶部に登録する登録データとして、前記複数の登録候補データの中から選択する
ことを特徴とする情報処理方法。
前記テストデータに対する前記類似度と、前記複数の登録候補データの各々の、前記データ記憶部に登録されている前記人物とは異なる複数の人物の生体情報を含む複数のデータの各々に対する類似度と、に基づいて前記登録データを選択する
ことを特徴とする付記11記載の情報処理方法。
前記複数の登録候補データのうち、前記テストデータに対する類似度が上位であり且つ前記複数のデータに対する前記類似度よりも高い登録候補データを、前記登録データとして選択する
ことを特徴とする付記12記載の情報処理方法。
前記登録データは、前記複数の登録候補データのうち、前記テストデータに対する類似度が最も高い登録候補データである
ことを特徴とする付記11乃至13の何れか1項に記載の情報処理方法。
前記複数の登録候補データの各々について、他の登録候補データを前記テストデータとして前記類似度を算出する
ことを特徴とする付記11乃至14の何れか1項に記載の情報処理方法。
前記複数の登録候補データの各々について、他の総ての登録候補データの各々を前記テストデータとして前記類似度を算出する
ことを特徴とする付記11乃至14の何れか1項に記載の情報処理方法。
前記複数の登録候補データのうち、他の登録候補データに対する類似度が最も高い登録候補データを、前記登録データとして選択する
ことを特徴とする付記16記載の情報処理方法。
前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアのスコア順位の和が最も小さい登録候補データである
ことを特徴とする付記17記載の情報処理方法。
前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアの和が最も大きい登録候補データである
ことを特徴とする付記17記載の情報処理方法。
コンピュータを、
同じ人物の生体情報を各々が含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出する手段、
前記複数の登録候補データのうち前記テストデータに対する類似度が上位の登録候補データを、前記人物の照合のためにデータ記憶部に登録する登録データとして、前記複数の登録候補データの中から選択する手段、
として機能させるプログラム。
付記20記載のプログラムを記録したコンピュータが読み取り可能な記録媒体。
110…画像取得部
120…照合スコア計算部
130…登録画像選択部
140…画像登録部
150…データ記憶部
160…照合部
200…プロセッサ
202…主記憶部
204…通信部
206…入出力インターフェース部
208…システムバス
210…出力装置
212…入力装置
214…記憶装置
300…撮像装置
400,700…記憶装置
500…情報処理装置
510…類似度計算部
520…照合部
530…登録データ選択部
600…データ取得装置
1000…認証システム
Claims (21)
- 同じ人物の生体情報を各々が含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出する類似度計算部と、
前記複数の登録候補データのうち前記テストデータに対する前記類似度が上位の登録候補データを、前記人物の照合のためにデータ記憶部に登録する登録データとして、前記複数の登録候補データの中から選択する登録データ選択部と
を有することを特徴とする情報処理装置。 - 前記データ記憶部は、前記人物とは異なる複数の人物の生体情報を含む複数のデータを有し、
前記登録データ選択部は、前記テストデータに対する前記類似度と、前記複数の登録候補データの各々の前記複数のデータの各々に対する類似度と、に基づいて前記登録データを選択する
ことを特徴とする請求項1記載の情報処理装置。 - 前記登録データ選択部は、前記複数の登録候補データのうち、前記テストデータに対する類似度が上位であり且つ前記複数のデータに対する前記類似度よりも高い登録候補データを、前記登録データとして選択する
ことを特徴とする請求項2記載の情報処理装置。 - 前記登録データは、前記複数の登録候補データのうち、前記テストデータに対する類似度が最も高い登録候補データである
ことを特徴とする請求項1乃至3の何れか1項に記載の情報処理装置。 - 前記類似度計算部は、前記複数の登録候補データの各々について、他の登録候補データを前記テストデータとして前記類似度を算出する
ことを特徴とする請求項1乃至4の何れか1項に記載の情報処理装置。 - 前記類似度計算部は、前記複数の登録候補データの各々について、他の総ての登録候補データの各々を前記テストデータとして前記類似度を算出する
ことを特徴とする請求項1乃至4の何れか1項に記載の情報処理装置。 - 前記登録データ選択部は、前記複数の登録候補データのうち、他の登録候補データに対する類似度が最も高い登録候補データを、前記登録データとして選択する
ことを特徴とする請求項6記載の情報処理装置。 - 前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアのスコア順位の和が最も小さい登録候補データである
ことを特徴とする請求項7記載の情報処理装置。 - 前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアの和が最も大きい登録候補データである
ことを特徴とする請求項7記載の情報処理装置。 - 人物の生体情報データを取得するデータ取得装置と、
複数の人物の複数の生体情報データが登録された記憶装置と、
前記データ取得装置が取得した前記生体情報データと前記記憶装置に登録されている前記複数の生体情報データとの類似度を算出する類似度計算部と、前記データ取得装置が取得した前記生体情報データが表す人物が前記記憶装置に登録されている人物であるか否かを前記類似度に基づいて判定する照合部と、を有する情報処理装置と、を有する照合システムであって、
前記類似度計算部は、同じ人物の生体情報を各々が含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出するように更に構成されており、
前記情報処理装置は、前記人物の照合のために前記記憶装置に登録する登録データを、前記複数の登録候補データの各々の、前記テストデータに対する前記類似度と、前記複数の人物の前記複数の生体情報データの各々に対する類似度と、に基づいて前記複数の登録候補データの中から選択する登録データ選択部を更に有する
ことを特徴とする照合システム。 - 同じ人物の生体情報を含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出し、
前記複数の登録候補データのうち前記テストデータに対する類似度が上位の登録候補データを、前記人物の照合のためにデータ記憶部に登録する登録データとして、前記複数の登録候補データの中から選択する
ことを特徴とする情報処理方法。 - 前記テストデータに対する前記類似度と、前記複数の登録候補データの各々の、前記データ記憶部に登録されている前記人物とは異なる複数の人物の生体情報を含む複数のデータの各々に対する類似度と、に基づいて前記登録データを選択する
ことを特徴とする請求項11記載の情報処理方法。 - 前記複数の登録候補データのうち、前記テストデータに対する類似度が上位であり且つ前記複数のデータに対する前記類似度よりも高い登録候補データを、前記登録データとして選択する
ことを特徴とする請求項12記載の情報処理方法。 - 前記登録データは、前記複数の登録候補データのうち、前記テストデータに対する類似度が最も高い登録候補データである
ことを特徴とする請求項11乃至13の何れか1項に記載の情報処理方法。 - 前記複数の登録候補データの各々について、他の登録候補データを前記テストデータとして前記類似度を算出する
ことを特徴とする請求項11乃至14の何れか1項に記載の情報処理方法。 - 前記複数の登録候補データの各々について、他の総ての登録候補データの各々を前記テストデータとして前記類似度を算出する
ことを特徴とする請求項11乃至14の何れか1項に記載の情報処理方法。 - 前記複数の登録候補データのうち、他の登録候補データに対する類似度が最も高い登録候補データを、前記登録データとして選択する
ことを特徴とする請求項16記載の情報処理方法。 - 前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアのスコア順位の和が最も小さい登録候補データである
ことを特徴とする請求項17記載の情報処理方法。 - 前記類似度は、前記テストデータに対する照合スコアであり、
前記他の登録候補データに対する類似度が最も高い登録候補データは、前記他の登録候補データに対する前記照合スコアの和が最も大きい登録候補データである
ことを特徴とする請求項17記載の情報処理方法。 - コンピュータを、
同じ人物の生体情報を各々が含む複数の登録候補データの各々について、前記人物の生体情報を含むテストデータに対する類似度を算出する手段、
前記複数の登録候補データのうち前記テストデータに対する類似度が上位の登録候補データを、前記人物の照合のためにデータ記憶部に登録する登録データとして、前記複数の登録候補データの中から選択する手段、
として機能させるプログラム。 - 請求項20記載のプログラムを記録したコンピュータが読み取り可能な記録媒体。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US17/637,977 US20240104178A1 (en) | 2021-02-18 | 2021-11-19 | Information processing apparatus, information processing method, matching system, program, and storage medium |
JP2023500539A JP7571857B2 (ja) | 2021-02-18 | 2021-11-19 | 情報処理装置、情報処理方法、照合システム、プログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021023824 | 2021-02-18 | ||
JP2021-023824 | 2021-02-18 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022176294A1 true WO2022176294A1 (ja) | 2022-08-25 |
Family
ID=82931357
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/042538 WO2022176294A1 (ja) | 2021-02-18 | 2021-11-19 | 情報処理装置、情報処理方法、照合システム、プログラム及び記録媒体 |
Country Status (3)
Country | Link |
---|---|
US (1) | US20240104178A1 (ja) |
JP (1) | JP7571857B2 (ja) |
WO (1) | WO2022176294A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2025037390A1 (ja) * | 2023-08-15 | 2025-02-20 | 日本電気株式会社 | 情報処理システム、情報処理方法、及び記録媒体 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000163574A (ja) * | 1998-11-30 | 2000-06-16 | Oki Electric Ind Co Ltd | データ識別方法および識別装置 |
JP2010009389A (ja) * | 2008-06-27 | 2010-01-14 | Toshiba Corp | 辞書情報登録装置、辞書情報登録方法、顔認証装置および入退場管理システム |
JP2012221464A (ja) * | 2011-04-14 | 2012-11-12 | Hitachi Solutions Ltd | 指静脈認証システム及び認証精度向上手法 |
JP2020003873A (ja) * | 2018-06-25 | 2020-01-09 | 株式会社日立製作所 | 生体認証プログラム、生体認証方法 |
-
2021
- 2021-11-19 US US17/637,977 patent/US20240104178A1/en active Pending
- 2021-11-19 JP JP2023500539A patent/JP7571857B2/ja active Active
- 2021-11-19 WO PCT/JP2021/042538 patent/WO2022176294A1/ja active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000163574A (ja) * | 1998-11-30 | 2000-06-16 | Oki Electric Ind Co Ltd | データ識別方法および識別装置 |
JP2010009389A (ja) * | 2008-06-27 | 2010-01-14 | Toshiba Corp | 辞書情報登録装置、辞書情報登録方法、顔認証装置および入退場管理システム |
JP2012221464A (ja) * | 2011-04-14 | 2012-11-12 | Hitachi Solutions Ltd | 指静脈認証システム及び認証精度向上手法 |
JP2020003873A (ja) * | 2018-06-25 | 2020-01-09 | 株式会社日立製作所 | 生体認証プログラム、生体認証方法 |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2025037390A1 (ja) * | 2023-08-15 | 2025-02-20 | 日本電気株式会社 | 情報処理システム、情報処理方法、及び記録媒体 |
Also Published As
Publication number | Publication date |
---|---|
US20240104178A1 (en) | 2024-03-28 |
JP7571857B2 (ja) | 2024-10-23 |
JPWO2022176294A1 (ja) | 2022-08-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP2523149B1 (en) | A method and system for association and decision fusion of multimodal inputs | |
JP3439359B2 (ja) | 個人識別方法、個人識別装置および記録媒体 | |
JP4937607B2 (ja) | 生体認証方法及び生体認証装置 | |
JP5375823B2 (ja) | 画像照合装置、画像照合方法および画像照合用プログラム | |
US8374401B2 (en) | Biometric authentication device, authentication accuracy evaluation device and biometric authentication method | |
JP2007219955A (ja) | 質問応答システム、質問応答処理方法及び質問応答プログラム | |
KR20120114564A (ko) | 대표 지문 템플릿 생성 장치 및 방법 | |
KR20100073845A (ko) | 음성 및 영상에 기반한 성별-연령 판별방법 및 그 장치 | |
CN112560710B (zh) | 一种用于构建指静脉识别系统的方法及指静脉识别系统 | |
JPWO2020050413A1 (ja) | 認証用顔画像候補判定装置、認証用顔画像候補判定方法、プログラム、および記録媒体 | |
CN118097755A (zh) | 一种基于yolo网络的智慧人脸身份识别方法 | |
JP5812505B2 (ja) | マルチモーダル情報に基づく人口学的分析方法及びシステム | |
WO2022176294A1 (ja) | 情報処理装置、情報処理方法、照合システム、プログラム及び記録媒体 | |
CN108875549A (zh) | 图像识别方法、装置、系统及计算机存储介质 | |
CN113158777A (zh) | 质量评分方法、质量评分模型的训练方法及相关装置 | |
CN109033321A (zh) | 一种图像与自然语言特征提取及基于关键词的语言指示图像分割方法 | |
Su et al. | Evaluation of rarity of fingerprints in forensics | |
JP2005141437A (ja) | パターン認識装置及びその方法 | |
CN112351047B (zh) | 基于双引擎的声纹身份认证方法、装置、设备及存储介质 | |
WO2012029150A1 (ja) | 生体認証システム、生体認証方法、及びプログラム | |
US9443443B2 (en) | Personalized studying path generating method in serious game | |
JPH11232459A (ja) | 個人認証用の認証対象原情報登録方法および個人認証装置 | |
CN110969095A (zh) | 一种信鸽的身份信息分析方法及装置 | |
JPH10261083A (ja) | 個人同定装置及び個人同定方法 | |
JP7318804B2 (ja) | 類似度計算機、認証システム、類似度計算方法、類似度計算プログラムおよび類似度計算プログラムの生成方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
WWE | Wipo information: entry into national phase |
Ref document number: 17637977 Country of ref document: US |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21926722 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2023500539 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 21926722 Country of ref document: EP Kind code of ref document: A1 |