JP6056398B2 - Biometric authentication apparatus, biometric authentication method, and biometric authentication computer program - Google Patents

Description

  The present invention relates to, for example, a biometric authentication apparatus, a biometric authentication method, and a biometric authentication computer program that perform biometric authentication based on biometric information represented in a biometric image.

  In recent years, biometric authentication technology has been developed for authenticating a user of an apparatus or system based on a biometric image representing biometric information such as a hand or finger vein pattern, a fingerprint, or a palm print. A biometric authentication device using such a biometric authentication technology acquires, for example, a biometric image representing the biometric information of a user who intends to use the biometric authentication device. Then, the biometric authentication device collates the input biometric information that is the biometric information of the user represented in the biometric image with the registered biometric information that is biometric information that is represented in the biometric image of the registered user registered in advance. If the biometric authentication device determines that the input biometric information and the registered biometric information match based on the result of the verification process, the biometric authentication device authenticates the user as a registered user having a legitimate authority. The biometric authentication device permits an authenticated user to use a device in which the biometric authentication device is incorporated or another device connected to the biometric authentication device.

Biometric authentication techniques include a so-called 1: 1 authentication method and a 1: N authentication method. In the 1: 1 authentication method, registered biometric information corresponding to the identification information is specified by inputting user identification information such as a user name or an identification number together with the input biometric information of the user. Then, the biometric authentication device collates the input biometric information with only the specified registered biometric information.
On the other hand, in the 1: N authentication method, the identification information of the user is not input to the biometric authentication device, and the registered biometric information corresponding to the user cannot be specified. For this reason, the biometric authentication device collates the input biometric information with all registered biometric information registered in advance. The biometric authentication device authenticates the user as a registered user corresponding to the registered biometric information that most closely matches the input biometric information. Therefore, as the number of registered users increases, that is, as the number of registered biometric information increases, the number of executions of collation processing also increases. Therefore, the time required for the biometric authentication process is inevitably increased. In particular, when biometric authentication technology is used in a large-scale system, the number of registered users may become very large. When such a large-scale system employs a server-client system, access to the server is concentrated to execute biometric authentication processing, the load on the server increases, and the user inputs biometric information. There is a possibility that the waiting time until the authentication result is obtained becomes long.

  However, from the viewpoint of ease of use by the user, it is preferable that the time required for the biometric authentication process is short. Therefore, it is preferable to limit registered biometric information used for collation with input biometric information. In order to limit the registered biometric information used for collation, for example, a feature quantity that can be extracted from a biometric image by a simple process and requires a small amount of calculation is used. Such a feature quantity used for limiting the registered biometric information used for collation is hereinafter referred to as a selection feature quantity for convenience.

  For example, when biometric information of a hand is used for biometric authentication processing, a technique has been proposed that uses the shape of a hand together with fingerprint information or palm print information (see, for example, Patent Document 1).

  Further, in order to speed up the biometric authentication process, a technique that uses the feature of the palm shape for matching has also been proposed (see, for example, Patent Document 2).

Japanese Patent Laid-Open No. 10-307919 JP 2002-15298 A

  However, in some cases, the biometric authentication device may not be able to accurately extract the selection feature amount from the biometric image. For example, when a feature amount related to the outer shape of the hand is used as the feature amount for selection, the hand is too large or the position of the hand with respect to the sensor that captures the hand is shifted, so that a part of the hand becomes a biological image It may not be reflected. In addition, when a hand is placed in the vicinity of the outer edge of the imaging range of the sensor, the distortion of the lens used for the sensor is so large that it cannot be ignored, the shape of the hand shown in the biological image is distorted by the influence of the distortion. If the ambient light is too bright, the difference between the luminance of the background area where the hand is not captured on the biological image and the luminance of the area where the hand is captured may be small, and the outline of the hand may not be clear. In these cases, since the biometric authentication device cannot accurately detect at least a part of the contour of the hand, there is a possibility that the selection feature amount cannot be accurately extracted. If the selection feature quantity is not accurately extracted, the biometric authentication device cannot select the registered user who is the user who seeks biometric authentication as a target for verification, or the selection feature quantity is essentially similar. There is a risk of selecting a registered user who should not exist as a verification target.

  Therefore, in the present specification, in the 1: N authentication method, even when a selection feature amount for selecting a registered user to be verified cannot be extracted, a registered user to be verified can be appropriately selected. An object is to provide a biometric authentication device.

  According to one embodiment, a biometric authentication device is provided. The biometric authentication apparatus includes a biometric information acquisition unit that generates a biometric image in which a part including biometric information of a user is captured, and a verification data generation unit that generates verification data representing characteristics of the biometric information from the biometric image. A feature extraction unit for selection that extracts a feature for selection representing the feature of the shape of the part from the biological image, and a feature of another shape of the part that has a predetermined correlation with the feature for selection from the biological image. An alternative feature amount extracting unit that extracts an alternative feature amount to be expressed and corrects the alternative feature amount according to the correlation to calculate an approximate value of the selection feature amount, and for selection based on the quality of the part in the biological image It is determined whether or not the feature quantity can be extracted, and if it is determined that the selection feature quantity can be extracted, the selection feature quantity extraction unit extracts the selection feature quantity, while the selection feature quantity is determined to be unextractable. In the alternative feature extraction unit A selection feature amount evaluation unit that calculates an approximate value of the selection feature amount, a storage unit that stores registered user verification data and selection feature amounts for each of a plurality of registered users, and a plurality of registrations Among the users, a selection unit for selecting a registered user having a selection feature amount similar to the selection feature amount of the user or an approximate value of the selection feature amount, and for verification of the selected registered user A collation unit for collating the data with the user's collation data;

The objects and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.
It should be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention as claimed.

  In the 1: N authentication method, the biometric authentication device disclosed in the present specification can be used to select a registered user to be verified even when a selection feature amount for selecting a registered user to be verified cannot be extracted. Can be selected appropriately.

1 is a schematic configuration diagram of a biometric authentication device according to a first embodiment. (A) is a schematic plan view of a biological information acquisition part, (b) is a schematic sectional side view of the biological information acquisition part in arrow AA 'of (a). It is a functional block diagram of the process part of a terminal. (A) is a figure showing an example of a living body image which can detect the width of a palm. (B)-(e) is a figure which shows the example of the biometric image which respectively cannot detect the width | variety of a palm, or cannot detect correctly. It is an operation | movement flowchart of the feature-value evaluation process for selection. The width of the palm, which is an example of the feature for selection, the width of the finger, which is an example of the alternative feature, and the midpoint of the base of the thumb and index finger and the midpoint of the base of the little finger and the ring finger, which is another example of the alternative feature It is a figure which shows these distances. It is a functional block diagram of the process part of an authentication server. It is an operation | movement flowchart of a biometrics authentication process. It is an operation | movement flowchart of a registration process. It is a schematic block diagram of the biometrics apparatus by 2nd Embodiment.

Hereinafter, biometric authentication devices according to various embodiments will be described with reference to the drawings.
When performing biometric authentication processing for a user in accordance with the 1: N authentication method, this biometric authentication device is used for selecting a feature of the shape of a part representing biometric information in order to select a registered user to be verified. Use features. However, if it is determined that the selection feature quantity cannot be detected based on the quality of the part in the biometric image, this biometric authentication device is able to select another part of the part having a predetermined correlation with the selection feature quantity. An alternative feature amount representing a shape feature is obtained. Then, the biometric authentication device obtains an approximate value of the selection feature amount by correcting the alternative feature amount according to the correlation with the selection feature amount, and uses the approximate value as the selection feature amount of the registered user. To select the registered user to be verified.

In this embodiment, the biometric authentication device uses a hand vein pattern as biometric information used for biometric authentication. However, the biometric information used for biometric authentication may be other biometric information included in the hand, such as a palm print.
Further, in this specification, the term “collation process” is used to indicate a process of calculating a degree of similarity representing a degree of similarity between a user's biometric information and a registered user's biometric information. In addition, the term “biometric authentication process” is used not only to indicate a verification process but also to indicate an entire authentication process including a process for determining whether to authenticate a user using a result of the verification process. The

  FIG. 1 shows a schematic configuration diagram of a biometric authentication apparatus. As shown in FIG. 1, the biometric authentication device 1 includes at least one terminal 2, an authentication server 3, and a data server 4. The terminal 2 and the authentication server 3 are connected to each other via a communication network 5 such as a public communication line or a local area network. The authentication server 3 and the data server 4 are connected via, for example, a specific communication line or another communication network independent of the communication network 5. Note that the data server 4 may also be connected to the authentication server 3 via the communication network 5.

  The terminal 2 generates a biological image obtained by photographing a part (in this embodiment, the user's hand) including the biological information of the user who is going to be authenticated. And the terminal 2 produces | generates the data for collation used for a collation process from the biometric image. Further, the terminal 2 determines whether or not the palm width can be detected as the selection feature amount representing the feature of the hand shape. If the terminal 2 determines that the palm width cannot be detected, the terminal 2 has a correlation with the palm width and obtains a feature amount of another part including biological information, for example, the width of the finger. The palm width is approximately obtained from the feature amount. Then, the terminal 2 transmits to the authentication server 3 the collation data and the palm width as the selection feature or an approximate value thereof.

  When the biometric authentication process is executed, the authentication server 3 uses registered features having a selection feature amount similar to the user's selection feature amount or its approximate value from among the registered users registered in the data server 4. A person is selected as a target for the verification process. In this example, the authentication server 3 selects a registered user having a palm width that is substantially the same as the palm width of the user as a target for collation processing. And the authentication server 3 performs a collation process using the collation data produced | generated from the biometric image showing the biometric information of the selected registered user, and the collation data produced | generated from the user's biometric image. The authentication server 3 notifies the terminal 2 as to whether or not the user has been authenticated as any registered user as a result of the verification process.

The terminal 2 captures a part including the biometric information of the user at the time of executing the biometric authentication process and registering the biometric information, and generates a biometric image showing the part. And the terminal 2 produces | generates the data for collation from the biometric image, and extracts the feature-value for selection, or calculates the approximate value. For this purpose, the terminal 2 includes a biological information acquisition unit 21, a communication unit 22, a storage unit 23, and a processing unit 24.
Furthermore, the terminal 2 has an input unit (not shown) such as a keyboard, a mouse, or a touch pad for a user to input identification information such as a command, data, or a user name of the user to the terminal 2. May be. Further, the terminal 2 displays a guidance message or a biometric authentication result for placing a finger at a position where the biometric information acquisition unit 21 can acquire a biometric image in which biometric information is appropriately captured, such as a liquid crystal display. You may have a display part (not shown).

  The biometric information acquisition unit 21 generates a biometric image in which the vein pattern of the palm of the user is shown.

FIG. 2A is a schematic plan view of the biological information acquisition unit 21, and FIG. 2B is a schematic side sectional view of the biological information acquisition unit 21 taken along the arrow AA ′ in FIG.
The biological information acquisition unit 21 includes an area sensor 211 that is sensitive to infrared light and in which a solid-state imaging device such as a charge coupled device is two-dimensionally arranged. The area sensor 211 is disposed at the bottom of a rectangular parallelepiped casing 212 made of resin or metal so that the sensor surface faces upward.
Furthermore, the biological information acquisition unit 21 includes an illumination light source 213 such as an infrared light emitting diode that emits infrared light to illuminate the user's hand. The side walls around the housing 212 are guide members for guiding the user's hand to a position where the area sensor 211 can shoot, and a part of the user's palm is placed on the upper end of the side wall. Can be placed. Further, a protrusion 214 for fixing a user's finger is formed on the upper side of the housing 212 on the back side of the area sensor 211. Then, an imaging optical system (not shown) included in the biological information acquisition unit 21 forms an image of a palm vein pattern placed on the guide member on the area sensor 211. Then, the area sensor 211 generates a biological image showing the palm vein pattern. The value of each pixel of the biological image is, for example, a value in the range of 0 to 255, and the pixel value increases as the luminance of the pixel increases. In the present embodiment, in the biometric image, a region where a vein is shown and a portion where a hand is not shown are black, and a region where a hand portion where no vein is shown is bright.
The biological information acquisition unit 21 passes the generated biological image to the processing unit 24.

  The communication unit 22 includes a communication interface circuit for connecting the terminal 2 to the communication network 5. Then, the communication unit 22 transmits the data received from the processing unit 24 to the authentication server 3 via the communication network 5. Further, the communication unit 22 may pass information representing the authentication result received from the authentication server 3 via the communication network 5 to the processing unit 24.

  The storage unit 23 includes, for example, a nonvolatile read-only semiconductor memory and a readable / writable volatile semiconductor memory. The storage unit 23 stores various programs and data used for controlling the terminal 2. In addition, the storage unit 23 stores a program used to generate collation data from the biometric image and extract a selection feature amount. Further, the storage unit 23 may temporarily store the biometric image, the verification data, and the selection feature amount.

  The processing unit 24 includes one or a plurality of processors and their peripheral circuits. Then, the processing unit 24 generates collation data from the biometric image at the time of executing the biometric authentication process and the registration process, and extracts a selection feature quantity or an approximate value of the selection feature quantity.

  FIG. 3 is a functional block diagram of the processing unit 24 of the terminal 2. The processing unit 24 includes a subject area detection unit 25, a collation data generation unit 26, a selection feature amount evaluation unit 27, a selection feature amount extraction unit 28, and an alternative feature amount extraction unit 29. Each of these units included in the processing unit 24 is a functional module implemented by a computer program executed on a processor included in the processing unit 24. Alternatively, these units included in the processing unit 24 may be implemented in the terminal 2 as firmware.

The subject region detection unit 25 detects a subject region that is a region in which a part including biological information is captured in the biological image from the biological image. In the present embodiment, the value of the pixel included in the background area, which is a set of pixels where nothing is captured, is lower than the value of the pixel included in the subject area. Therefore, the subject area detection unit 25 detects, for example, a pixel having a pixel value equal to or greater than a predetermined threshold, and sets the detected set of pixels as the subject area. The predetermined threshold is set to 10, for example.
The subject area detection unit 25 stores information representing the subject area in the storage unit 23. The information representing the subject area can be, for example, a binary image having the same size as the size of the biological image. In this binary image, the value of the pixel included in the subject area is, for example, “1”. The value of the pixel set and included in the background area is set to “0”, for example.

  The verification data generation unit 26 generates verification data used in the verification process from the biological image. For example, when the authentication server 3 performs a matching process by minutia matching, the collation data generation unit 26 extracts feature points (referred to as minutia) used in the minutia matching from the biological image. The collation data generation unit 26 uses the position of the feature point or the type of the feature point as collation data.

  For this purpose, the collation data generation unit 26 distinguishes between a vein region that is a set of pixels in which the veins are shown and a non-vein region that is a set of pixels in which the veins are not shown in the subject region. In this embodiment, the value of the pixel in which the vein is shown is lower than the value of the pixel in which the vein is not shown. Therefore, the verification data generation unit 26 uses, for example, a set of pixels having a pixel value equal to or lower than the second binarization threshold in the subject area as a vein area, and has a pixel value larger than the second binarization threshold. The subject area is binarized so that the set of pixels is a non-vein area. The second binarization threshold is determined, for example, by discriminating and analyzing a preset fixed value (for example, 150), an average pixel value of each pixel in the subject area, or a pixel value included in the subject area. Is set to the threshold value.

  Next, for example, the collation data generation unit 26 performs a thinning process on a set of pixels having pixel values corresponding to veins on a binarized subject area, thereby thinning the veins. A binarized binary image is generated. The collation data generation unit 26 scans the thinned binary image using a plurality of templates corresponding to either the branching points or the end points of the veins, thereby thinning 2 when matching with any template. The position on the value image is detected. Then, the verification data generation unit 26 extracts the central pixel at the detected position as a minutiae.

  The verification data generation unit 26 may extract the minutia from the biological image by using another known method for obtaining the end point and the branch point of the vein as the minutia. In addition, the matching data generation unit 26 may obtain other feature quantities representing the features of the vein pattern as matching data on the biological image. For example, the verification data generation unit 26 may divide the subject region into a plurality of blocks and obtain the number of veins for each block as the verification data.

  Further, when the biometric authentication device 1 performs the matching process by pattern matching, the matching data generation unit 26 cuts out the biometric image itself or a part of the region including the vein region on the biometric image from the biometric image, and performs matching. It may be data. Furthermore, the verification data generation unit 26 may use, as the verification data, an image obtained by performing edge enhancement processing or unevenness correction processing on the entire biological image or a part of the biological image.

  The verification data generation unit 26 stores the verification data in the storage unit 23.

  The selection feature quantity evaluation unit 27 evaluates whether or not a selection feature quantity used to select a registered user to be verified from a plurality of registered users can be extracted.

  In the present embodiment, the biometric authentication device 1 uses the width of the palm as the selection feature amount, particularly in this example, between the contour of the hand along the thumb side of the index finger and the contour of the hand along the outside of the little finger. The distance is the palm width. The reason for using the palm width as the selection feature amount is as follows.

As long as the palm width varies to some extent and the distance from the area sensor is constant, the biometric authentication device 1 appropriately selects the registered user to be collated, since there is little fluctuation depending on the shooting environment. it can.
In addition, as a feature quantity having characteristics similar to the palm width, there is a finger length or a finger width. However, in this embodiment, since the tip of the finger is out of the imaging range of the biometric information acquisition unit 21, the biometric authentication device 1 cannot obtain the length of the user's finger from the biometric image. Also, since the finger width size is smaller than the palm width size, if the calculated finger width includes an error of one to several pixels, the calculated error of the palm will have the same error. The impact is greater than if included. Therefore, the palm width is more preferable as the selection feature amount than the finger length and finger width.

  FIG. 4A is a diagram illustrating an example of a biological image in which the palm width can be detected. FIG. 4B to FIG. 4E are diagrams illustrating examples of biological images in which the palm width cannot be detected or cannot be detected accurately.

  In the biological image 400 shown in FIG. 4A, the range from the outside of the little finger to the contour on the thumb side of the index finger is shown. Therefore, for this biometric image 400, the biometric authentication device 1 can detect the palm width indicated by the arrow 401.

On the other hand, in the biological image 410 shown in FIG. 4B, the palm of the user's hand and the index finger are in contact with the inner frame 411 of the guide, and thus a part of the palm is not shown. Therefore, for this biometric image 410, the biometric authentication device 1 cannot detect the palm width.
Also, in the biological image 420 shown in FIG. 4C, a part of the palm is not shown because the index finger side of the user's hand is in contact with the inner frame 421 of the guide. Therefore, the biometric authentication device 1 cannot detect the palm width of the biometric image 420 as well.

Furthermore, in the biological image 430 shown in FIG. 4D, the peripheral portion of the user's hand is out of the allowable range 431 for reducing the peripheral light amount by the imaging optical system (not shown) of the biological information acquisition unit 21. ing. Therefore, there is a possibility that the biometric authentication device 1 cannot detect the palm width of the biometric image 430.
In addition, the biological image 440 shown in FIG. 4E is detected because the difference between the luminance value of the subject area and the luminance value of the background area in the vicinity of the contour of the user's hand is small due to the influence of ambient light. The contour of the subject area 441 is different from the original contour. Therefore, the biometric authentication device 1 may not be able to detect the palm width of the biometric image 440.

  As described above, if the quality of the part including the biometric information that appears in the biometric image is deteriorated for various reasons, the biometric authentication device 1 may not be able to detect the palm width that is the selection feature amount. is there. Therefore, the selection feature quantity evaluation unit 27 checks whether or not the deterioration condition of the quality of the part including the biological information on the biological image as described above is satisfied, and determines that the biological image satisfies the deterioration condition. In this case, an alternative feature amount is extracted. For example, the selection feature quantity evaluation unit 27 extracts a substitute feature quantity if a part of the part including the biological information related to the selection feature quantity cannot be detected on the biological image.

  FIG. 5 is an operation flowchart of the selection feature quantity evaluation process executed by the selection feature quantity evaluation unit 27.

  The selection feature amount evaluation unit 27 detects the contour of the subject area on the biological image (step S101). For example, the selection feature quantity evaluation unit 27 sets a pixel in which any of the adjacent pixels is included in the background region among the pixels included in the subject region as a pixel on the contour of the subject region.

  The selection feature amount evaluating unit 27 determines whether or not each of the left end contour and the right end contour of the subject area matches the inner frame of the guide by a predetermined ratio or more (step S102). Since the position of the guide with respect to the area sensor and the optical system of the biological information acquisition unit 21 is fixed, the pixel in which the inner frame of the guide appears on the biological image is known in advance. Therefore, for example, the coordinates of each of the plurality of pixels in which the inner frame of the guide appears on the biological image are stored in the storage unit 23 in advance. Then, when the selection feature quantity evaluating unit 27 matches the coordinates of the inner frame of the guide with respect to the leftmost pixel of the contour and the rightmost pixel of the contour of each row of the biometric image, the pixel is defined as the inner frame of the guide. It is determined that they match. The predetermined ratio can be, for example, 90% of the length of the contour at the left end of the subject area or the length of the contour at the right end of the subject area.

  When both the left edge outline and the right edge outline of the subject area coincide with the inner frame of the guide by a predetermined ratio or more (Yes in step S102), the user's hand is too large and the user's hand outline Is estimated not to appear in the biological image. Therefore, the selection feature quantity evaluation unit 27 causes the alternative feature quantity extraction unit 29 to obtain an alternative feature quantity (step S103).

  On the other hand, when at least one of the left edge outline and the right edge outline of the subject area matches the inner frame of the guide is less than a predetermined ratio (step S102-No), the selection feature amount evaluation unit 27 Check if the palm is not entirely shifted to the left or right. For this purpose, the selection feature quantity evaluation unit 27 determines whether the ratio of the outline that has a higher proportion matching the inner frame of the guide out of the left end contour and the right end contour of the subject area is equal to or greater than a predetermined ratio. (Step S104). When the coincidence ratio is equal to or higher than the predetermined ratio (step S <b> 104 -Yes), the selection feature amount evaluation unit 27 determines that the position of the user's hand with respect to the biometric information acquisition unit 21 is shifted left and right. Then, the selection feature quantity evaluation unit 27 displays a guidance message that prompts the user to align the center position of the hand with the center of the biometric information acquisition unit 21 via the display unit (not shown) of the terminal 2 (step S105). . Thereafter, subject area detection and selection feature amount evaluation processing is performed on the biological image regenerated by the biological information acquisition unit 21. In step S104, the selection feature amount evaluation unit 27 determines that the position of the user's hand is shifted when the distance between the other contour and the inner frame of the guide is a predetermined distance or more. You may add as conditions for. In this case, the predetermined distance can be, for example, the number of pixels corresponding to 1 cm in actual distance.

In step S104, when the ratio that matches the inner frame of the guide is less than the predetermined ratio (step S104-No), at least the relevant part of the contour of the subject area used for calculating the palm width is a living body. Presumed to be in the image.
Therefore, next, the selection feature quantity evaluation unit 27 determines whether or not a portion of the outline of the subject region that is used for calculating the palm width is out of a preset allowable range (step S106). . The allowable range is a range on the biological image that is such that the distortion of the optical system of the biological information acquisition unit 21 is negligible (for example, the amount of deviation due to distortion is equal to or less than the number of pixels corresponding to 1 mm in actual size). can do. Alternatively, the allowable range may be a range on the biological image in which the peripheral light amount ratio for the optical system of the biological information acquisition unit 21 is a predetermined ratio (for example, 50%) or more that does not affect the detection of the hand contour. This allowable range is determined according to the performance of the optical system, and can be set in advance. Therefore, information indicating the allowable range is stored in the storage unit 23 in advance.

  In this embodiment, the part used for calculating the palm width is the contour of the hand near the base of the little finger along the outside of the little finger and the hand near the base of the index finger along the thumb side of the index finger. Is the outline. The contour of the hand near the base of the little finger along the outside of the little finger is, for example, when the little finger is positioned on the right end side of the biological image, It can be a portion included in the range of 1/3 to 1/2 of the length in the vertical direction. Also, since the thumb does not touch the upper end of the biometric image, the left end of the contour of the subject area near the upper end of the biometric image is an outline along the thumb side of the index finger. The midpoint between the base of the index finger and the base of the thumb is located on the lowermost position on the biometric image in the contour along the thumb side of the index finger. That is, from the pixel located at the left end of the contour of the subject area in the vicinity of the upper end of the biometric image to the pixel located at the lowermost of the pixels on the contour continuous with the pixel, the contour along the thumb side of the index finger Become. Therefore, the contour of the hand near the base of the index finger along the thumb side of the index finger is, for example, a pixel of the contour of the subject area corresponding to the midpoint between the base of the index finger and the base of the thumb, and a biological image from the bottom The contour portion is included in the range of 1/3 to 1/2 of the distance to the upper end of the.

  When at least a part of the portion used for calculating the palm width is out of the allowable range (step S106: Yes), the selection feature value evaluation unit 27 may not be able to accurately calculate the palm width. Is determined. Therefore, the selection feature quantity evaluation unit 27 causes the alternative feature quantity extraction unit 29 to obtain an alternative feature quantity (step S103).

  On the other hand, when the whole part used for calculation of the palm width is included in the allowable range (step S106-No), the part used for extracting the feature value for selection in the outline of the palm is on the biological image. Presumed to be in the proper position. Therefore, next, the selection feature amount evaluating unit 27 calculates the average value of the luminance values of the pixels on the contour included in the portion used for calculating the palm width and the pixels in the background region adjacent to the pixels on the contour. The difference between the luminance value and the average value is obtained. Then, the selection feature amount evaluation unit 27 determines whether or not the difference between the average values of the luminance values (hereinafter referred to as the contour luminance difference for convenience) is less than a predetermined luminance threshold (step S107).

  If the contour brightness difference is less than the predetermined brightness threshold (step S107—Yes), the selection feature amount evaluation unit 27 cannot accurately detect the contour of the subject region, and can accurately determine the palm width. It is determined that there is a possibility that it cannot be performed. Therefore, the selection feature quantity evaluation unit 27 causes the alternative feature quantity extraction unit 29 to obtain an alternative feature quantity (step S103).

On the other hand, when the contour luminance difference is equal to or larger than the predetermined luminance threshold (step S107—No), the selection feature amount evaluation unit 27 determines that the palm width can be accurately obtained. Therefore, the selection feature quantity evaluation unit 27 causes the selection feature quantity extraction unit 28 to obtain the palm width, which is the selection feature quantity (step S108). Note that the luminance threshold is set to, for example, 10% to 20% of the average value of the luminance values of the pixels on the contour of the subject area.
After step S103, S105, or S108, the selection feature quantity evaluation unit 27 ends the selection feature quantity evaluation process.
Note that the selection feature amount evaluation unit 27 may change the order of determination in steps S102, S104, S106, and S107.

  Hereinafter, the selection feature amount and the alternative feature amount will be described. FIG. 6 illustrates the width of the palm, which is an example of the feature for selection, the width of the finger, which is an example of the alternative feature, and the midpoint of the base of the thumb and index finger, and the base of the little finger, the ring finger, which is another example of the alternative feature. It is a figure which shows the distance between the middle points. In the present embodiment, the selection feature amount is the width of the palm, and as indicated by an arrow 601, along the outer side of the little finger, the outline of the hand near the base of the little finger and the thumb side of the index finger, It is expressed as the distance from the contour of the hand near the base of the index finger. On the other hand, an example of the alternative feature amount is the width of any finger as indicated by an arrow 602. Another example of the alternative feature amount is a distance between the midpoint of the base of the thumb and the index finger and the midpoint of the base of the ring finger and the little finger, as indicated by an arrow 603.

  When the selection feature quantity evaluation unit 27 determines that the selection feature quantity can be extracted from the biometric image, the selection feature quantity extraction unit 28 selects the feature quantity for the part including the biometric information as a selection feature quantity. Find the width of the palm in the biological image.

  For example, when the little finger is positioned on the right end side of the biological image, that is, when the left hand is reflected in the biological image, the contour of the right end of the subject area corresponds to the contour of the hand along the outside of the little finger. Further, since the tip of the thumb does not reach the upper end of the biometric image, the pixel on the contour at the left end of the subject region and the pixel on the contour continuous with the pixel near the upper end of the biometric image are along the thumb side of the index finger. Corresponds to the contour of the hand. When the right hand is shown in the biological image, the contour of the left end of the subject region corresponds to the contour of the hand along the outside of the little finger, and the pixel on the right contour of the subject region near the upper end of the biological image and its The pixel on the contour that is continuous with the pixel corresponds to the contour of the hand along the thumb side of the index finger.

The selection feature amount extraction unit 28 uses an arbitrary pixel on the contour of the hand along the outside of the little finger as an end point on the little finger side, and an arbitrary pixel on the contour of the hand along the thumb side of the index finger as an index finger side end point. . Then, the selection feature amount extraction unit 28 calculates the distance between the two points for each pair while changing the pair of the little finger end point and the index finger end point in various ways. Then, the selection feature quantity extraction unit 28 sets the minimum value of these distances as the palm width.
The selection feature amount extraction unit 28 stores the palm width in the storage unit 23.

  When the selection feature quantity evaluation unit 27 determines that there is a possibility that the selection feature quantity cannot be obtained from the biological image, the alternative feature quantity extraction unit 29 has a predetermined correlation with the selection feature quantity. And an alternative feature amount representing a feature of another shape of the part including the biological information. Then, the alternative feature quantity extraction unit 29 corrects the alternative feature quantity according to the correlation to obtain an approximate value of the selection feature quantity.

  In the present embodiment, as described above, the alternative feature amount extraction unit 29 has a high correlation with the width of the palm, the width of any finger, or the midpoint of the base of the thumb and index finger, the ring finger, and the little finger. The distance between the midpoints of the roots of the is obtained as an alternative feature amount. For this purpose, the alternative feature amount extraction unit 29 detects the base of the finger. For example, the alternative feature amount extraction unit 29 sets a plurality of horizontal scanning lines at equal intervals with respect to the vertical direction of the biological image. Then, the alternative feature amount extraction unit 29 detects a contour pixel on the subject area on each scanning line, so that a background area sandwiched between two adjacent subject areas is set as an area between two adjacent fingers. To detect. Then, the substitute feature amount extraction unit 29 detects an area between the fingers for each scanning line in order from the upper end side of the biological image, and the lower end of the area between the fingers is located in the root of two adjacent fingers. Let it be a point. Further, when the little finger is located on the right end side of the biological image, the alternative feature amount extraction unit 29 sets the rightmost midpoint among the detected midpoints of the two finger roots to the middle of the little finger and the ring finger. Let it be a point. Similarly, the alternative feature amount extraction unit 29 sets the leftmost midpoint of the detected midpoints of the two finger bases as the midpoint of the base of the index finger and the middle finger. When the little finger is positioned on the left end side of the biological image, the substitute feature amount extraction unit 29 conversely selects the leftmost midpoint among the detected midpoints of the two finger bases as the little finger. The midpoint of the base of the ring finger is the midpoint of the index finger and the middle finger.

Further, the alternative feature amount extraction unit 29 finds a straight line (for example, a line 621 in FIG. 6) that approximately represents the contour of the middle finger side of the index finger in order to detect the midpoint of the base of the index finger and the thumb. Then, the alternative feature amount extraction unit 29 determines a position where a straight line (for example, a line 622 in FIG. 6) that is orthogonal to the straight line and passes through the middle point between the index finger and the middle finger intersects the contour on the thumb side of the index finger. The midpoint of the base of the index finger and thumb.
Note that the alternative feature amount extraction unit 29 is the second pixel from the left end of the biometric image in the vicinity of the upper end of the biometric image when the little finger is positioned on the right end side of the biometric image. A plurality of pixels corresponding to the pixels on the contour are detected. The alternative feature amount extraction unit 29 can detect a straight line that approximately represents the middle finger side contour of the index finger by performing Hough transform on the pixels on the plurality of contours.

  The alternative feature amount extraction unit 29 can detect three midpoints of the bases of two fingers from the biometric image, and can detect the midpoints of the bases of the index and thumbs. The distance between the middle point of the base of the point, little finger, and ring finger is calculated as an alternative feature amount. This distance corresponds to the width of three fingers (index finger, middle finger, and ring finger). On the other hand, in the present embodiment, the width of the palm corresponds to the width of four fingers (index finger, middle finger, ring finger, little finger). Therefore, the alternative feature amount extraction unit 29 calculates an approximate value of the palm width by multiplying this distance by a correction coefficient having a value of about 1.3 to 1.4.

  Alternatively, the alternative feature amount extraction unit 29 may calculate the distance between the midpoints of the two adjacent finger bases as the width of one finger. In this case, the alternative feature amount extraction unit 29 calculates an approximate value of the palm width by multiplying the width of the finger by a correction coefficient having a value of about 4.0 to 5.0.

  On the other hand, if only two or less midpoints of the bases of adjacent fingers are detected, it is estimated that the user's hand is shown in the biological image with at least some of the fingers closed. Therefore, there is a possibility that either the midpoint of the base of the little finger and the ring finger or the midpoint of the base of the middle finger and the index finger is not detected. Therefore, in this case, the alternative feature quantity extraction unit 29 calculates the distance between the midpoints of the two adjacent finger bases. The alternative feature amount extraction unit 29 sets the shortest of these distances as the finger width. Alternatively, the alternative feature amount extraction unit 29 may compare these distances with a preset reference value of the finger width, and may use the distance closest to the reference value as the finger width. Alternatively, in the alternative feature amount extraction unit 29, any one of the distances between the midpoints of the two adjacent finger bases is about 0.9 to 1.1 times the double of the finger width reference value. If it is, 1/2 of the distance may be used as the finger width.

In addition, when the selection feature amount evaluation unit 27 determines that both the left end outline and the right end outline of the subject region coincide with the inner frame of the guide by a predetermined ratio or more, the selection feature amount evaluation unit 27 is more likely to protrude from the guide. It is estimated that the user's hand is large. Therefore, in this case, the alternative feature quantity extraction unit 29 calculates the distance between the inner frame on the left end side of the guide and the inner frame on the right end side of the guide as an alternative feature quantity. In this case, the alternative feature amount extraction unit 29 uses the distance between the inner frame on the left end side of the guide and the inner frame on the right end side of the guide as it is as an approximate value of the palm width.
The alternative feature quantity extraction unit 29 stores the approximate value of the selection feature quantity in the storage unit 23.

When executing the registration process, the processing unit 24 transmits the verification data and the selection feature quantity or the approximate value of the selection feature quantity to the authentication server 3 via the communication unit 22 together with the user identification information.
In addition, when the biometric authentication process is executed, the processing unit 24 transmits the verification data and the selection feature quantity or the approximate value of the selection feature quantity to the authentication server 3 via the communication unit 22.

  The data server 4 stores data used for biometric authentication processing for all registered users. For this purpose, the data server 4 includes a communication unit 41, a storage unit 42, and a control unit 43.

  The communication unit 41 includes an interface circuit for connecting the data server 4 and the authentication server 3 so that they can communicate with each other. The communication unit 41 receives a registration request signal from the authentication server 3 and passes it to the control unit 43 when executing the registration process. Further, the communication unit 41 transmits the registered user verification data and the selection feature quantity stored in the storage unit 42 to the authentication server 3 when the biometric authentication process is executed.

  The storage unit 42 includes, for example, at least one of a semiconductor memory, a magnetic disk device, and an optical disk device. The storage unit 42 stores various programs and data used for controlling the data server 4. The storage unit 42 stores user identification information, collation data, and selection feature amounts for all registered users.

  The control unit 43 includes at least one processor and peripheral circuits. When receiving a registration request signal from the authentication server 3 during execution of the registration process, the control unit 43 stores various information of the registered user included in the registration request signal in the storage unit 42. When the control unit 43 receives a selection feature quantity request signal from the authentication server 3 during execution of the biometric authentication process, the control unit 43 reads the selection feature quantity of each registered user from the storage unit 42 and transmits it to the authentication server 3. . When receiving the verification data request signal from the authentication server 3, the control unit 43 reads the verification data of the selected registered user from the storage unit 42 and transmits it to the authentication server 3.

At the time of executing the registration process, the authentication server 3 transmits the verification data and the selection feature quantity received from the terminal 2 to the data server 4 together with the user identification information. If an approximate value of the feature value for selection is obtained instead of the feature value for selection, the authentication server 3 transmits the approximate value to the data server 4 as the feature value for selection.
Further, when executing the biometric authentication process, the authentication server 3 selects a registered user to be a target of the matching process based on the selection feature amount received from the terminal 2. Then, the authentication server 3 executes a verification process using the selected registered user verification data and the verification data received from the terminal 2, and determines whether to authenticate the user as any registered user. To do. When the authentication server 3 authenticates the user as one of the registered users, the authentication server 3 permits the user to use the device on which the authentication server 3 is mounted. Alternatively, the authentication server 3 transmits a signal indicating that the user has been authenticated to another device (not shown) and permits the user to use the other device.
For this purpose, the authentication server 3 includes a communication unit 31, a storage unit 32, and a processing unit 33.

The communication unit 31 includes a communication interface circuit for connecting the authentication server 3 to the communication network 5. Then, the communication unit 31 passes information received from the terminal 2 via the communication network 5 to the processing unit 33. The communication unit 31 may transmit information representing the authentication result received from the processing unit 33 to the terminal 2 via the communication network 5.
Furthermore, the communication unit 31 includes an interface circuit for connecting the authentication server 3 and the data server 4 so that they can communicate with each other. When the biometric authentication process is executed, the communication unit 31 receives the identification information and selection feature amount of each registered user and the selected registration user verification data from the data server 4 and passes them to the processing unit 33. . Further, when executing the registration process, the communication unit 31 transmits the matching data and the selection feature amount received from the terminal 2 to the data server 4 together with the user identification information.

  The storage unit 32 includes, for example, at least one of a semiconductor memory, a magnetic disk device, and an optical disk device. The storage unit 32 stores various programs and data used for controlling the authentication server 3. The storage unit 32 stores identification information of all registered users. Furthermore, when the biometric authentication process is executed, the storage unit 32 temporarily stores the identification information of each registered user, the selection feature amount, and the verification data of the selected registered user. Further, when executing the registration process, the storage unit 32 temporarily stores the matching data, the selection feature quantity, or the approximate value of the selection feature quantity received from the terminal 2 together with the user identification information.

  The processing unit 33 includes at least one processor and peripheral circuits. And the process part 33 performs the collation with the selection of a registered user and a user's input biometric information, and the registration biometric information of the selected registered user in biometrics authentication processing. The processing unit 33 executes registration processing.

  FIG. 7 is a functional block diagram of the processing unit 33 of the authentication server 3. The processing unit 33 includes a selection unit 34, a collation unit 35, an authentication determination unit 36, and a registration unit 37. Each of these units included in the processing unit 33 is a functional module implemented by a computer program executed on a processor included in the processing unit 33. Or these each part which the process part 33 has may be mounted in the authentication server 3 as firmware.

  The selection unit 34, the verification unit 35, and the authentication determination unit 36 are used when the biometric authentication process is executed, while the registration unit 37 is used when the registration process is executed. Therefore, in the following, the biometric authentication process and the registration process will be described separately.

(Biometric authentication process)
First, a biometric authentication process for determining whether to authenticate a user as any registered user will be described.
When the biometric authentication process is executed, the processing unit 33 receives the selection feature quantity of the user or an approximate value of the selection feature quantity from the terminal 2, and the selection feature is sent to the data server 4 via the communication unit 31. Send a quantity request signal. Then, the processing unit 33 receives selection feature amounts of all registered users from the data server 4 and temporarily stores them in the storage unit 32. In the above description, the authentication server 3 obtains the selection feature quantity from the data server 4 when the biometric authentication process is executed. However, the authentication server 3 obtains the selection feature quantity in advance and stores it in the storage unit 32. You may keep it.

  The selection unit 34 selects a registered user to be collated based on the selection feature amount. For this purpose, the selection unit 34 receives the user's biometric information based on the selection feature quantity of the user or the approximate value of the selection feature quantity received from the terminal 2 and the selection feature quantity of each registered user. A selection score representing the degree of similarity between biometric information of registered users is calculated. Note that the amount of computation required to calculate the selection score for one registered user is smaller than the amount of computation for the matching process for one registered user. Therefore, when the selection unit 34 selects a registered user to be verified based on the selection score, the biometric authentication device 1 reduces the amount of calculation compared to executing verification processing for all registered users. it can.

  For example, the selection unit 34 calculates the absolute value of the difference between the selection feature quantity of the user or the approximate value of the selection feature quantity and the selection feature quantity of the registered user as the selection score. That is, as the selection score is smaller, the user's selection feature quantity or the approximate value of the selection feature quantity and the registered user's selection feature quantity are more similar.

  The selection unit 34 calculates a selection score for each registered user. Then, the selection unit 34 selects the second predetermined number of registered users in ascending order of the selection score, and selects the selection feature amount similar to the selection feature amount of the user or the approximate value of the selection feature amount. Select as a registered user. The second predetermined number is a value smaller than the first predetermined number, which is the total number of registered users, and is set in advance according to the processing capability of the authentication server 3.

  Alternatively, the selection unit 34 uses only registered users whose selection score is less than a predetermined value for registered use having a selection feature amount similar to the user's selection feature amount or an approximate value of the selection feature amount. You may choose as a person. As a result, the number of selected registered users is less than the second predetermined number, and as a result, there is a possibility that the calculation amount of the entire biometric authentication process can be reduced. The predetermined value is set such that, for example, the principal rejection rate due to the fact that the registered user is not selected is equal to or less than the allowable maximum value experimentally.

  The selection unit 34 returns the identification information of the selected registered user to the processing unit 33. Then, the processing unit 33 transmits the identification information of the selected registered user to the data server 4 together with the verification data request signal. Then, the processing unit 33 receives the selected registered user verification data from the data server 4 and temporarily stores it in the storage unit 32.

  Using the user verification data received from the terminal 2 and the selected registered user verification data, the verification unit 35 uses the input biometric information of the user and the registered biometric data of each selected registered user. Match information. And the collation part 35 calculates | requires the similarity showing the degree to which input biometric information is similar to registration biometric information for every selected registration user as a result of collation processing.

  The matching unit 35 can use, for example, minutia matching or pattern matching as the matching process.

When the minutia matching is used, the collation unit 35 obtains the number of minutiaes that match between the minutia included in the registered user's collation data for the vein pattern and the minutia included in the user's collation data. Then, the collation unit 35 can calculate the similarity between the user's vein pattern and the registered user's vein pattern by dividing the matching number by the number of minutiae extracted for the user's vein pattern.
Alternatively, when using pattern matching, the matching unit 35 calculates the normalized cross-correlation value while changing the relative position between the biometric image of the user and the biometric image of the registered user. Then, the matching unit 35 sets the maximum value of the normalized cross-correlation value as the similarity between the user's vein pattern and the registered user's vein pattern.

  The collation unit 35 obtains the maximum value among the similarities calculated for each of the selected registered users. Then, the collation unit 35 passes the maximum value of the similarity and the identification information of the registered user corresponding to the maximum value to the authentication determination unit 36.

  The authentication determination unit determines that the input biometric information matches the registered biometric information corresponding to the maximum value of the similarity when the maximum value of the similarity is equal to or greater than the authentication determination threshold. The authentication determination unit 36 authenticates the user as a registered user having registered biometric information corresponding to the maximum value of the similarity. When authenticating the user, the authentication determination unit 36 notifies the processing unit 33 of the authentication result.

  On the other hand, the authentication determination unit 36 determines that the input biometric information and the registered biometric information do not match when the maximum similarity is less than the authentication determination threshold. In this case, the authentication determination unit 36 does not authenticate the user. The authentication determination unit 36 notifies the processing unit 33 of an authentication result indicating that the user authentication has failed. Then, the processing unit 33 may transmit authentication result information representing the authentication result to the terminal 2.

  The authentication determination threshold is preferably set to a value such that the authentication determination unit 36 succeeds in authentication only when any registered user is a user. The authentication determination threshold is preferably set to a value that causes the authentication determination unit 36 to fail authentication when another person different from the registered user is the user. For example, the authentication determination threshold value may be a value obtained by adding a value obtained by multiplying the difference between the maximum value and the minimum value of similarity by 0.7 to the minimum value of similarity.

FIG. 8 is an operation flowchart of the biometric authentication process.
The processing unit 24 of the terminal 2 acquires a biological image in which the user's biological information is captured from the biological information acquisition unit 21 (step S201). Then, the subject region detection unit 25 of the processing unit 24 detects a subject region in which a part including biological information is shown on the biological image (step S202). Thereafter, the collation data generation unit 26 of the processing unit 24 generates collation data for the biological information from the subject region (step S203).

  Further, the selection feature quantity evaluation unit 27 of the processing unit 24 determines whether or not the selection feature quantity can be extracted from the biological image in accordance with the operation flowchart shown in FIG. 5 (step S204). When the selection feature quantity evaluation unit 27 determines that the biometric information needs to be reacquired, the processing unit 24 repeats the processes after step S201.

On the other hand, when the selection feature quantity evaluation unit 27 determines that the selection feature quantity can be extracted, the selection feature quantity extraction unit 28 of the processing unit 24 obtains the selection feature quantity from the subject region (step S205).
Alternatively, if the selection feature quantity evaluation unit 27 determines that there is a possibility that the selection feature quantity may not be extracted, the alternative feature quantity extraction unit 29 of the processing unit 24 extracts the alternative feature quantity from the subject region and replaces it. Based on the feature value, an approximate value of the feature value for selection is obtained (step S206).

  After step S <b> 205 or S <b> 206, the processing unit 24 transmits the matching data and the selection feature quantity or the approximate value of the selection feature quantity to the authentication server 3 via the communication network 5.

Upon receiving the user verification data and the selection feature quantity or the approximate value of the selection feature quantity from the terminal 2, the authentication server 3 transmits a selection feature quantity request signal to the data server 4. Then, the authentication server 3 acquires feature values for selection of all registered users from the data server 4 (step S207). The selection unit 34 of the processing unit 33 of the authentication server 3 obtains the selection scores of all registered users based on the selection feature amount of the user or its approximate value and the selection feature amount of each registered user. Calculate (step S208). The selection unit 34 sequentially starts with the registered users having a selection score indicating that the degree of similarity between the user's selection feature quantity or its approximate value and the registration user's selection feature quantity is high. A predetermined number of registered users of 2 is selected (step S209).
The authentication server 3 transmits the identification information of the selected registered user and a verification data request signal to the data server 4. Then, the authentication server 3 receives the verification data of the selected registered user from the data server 4.

  For each selected registered user, the collation unit 35 determines the similarity between the user's biometric information and the registered user's biometric information based on the user's collation data and the registered user's collation data. Each is calculated (step S210). Then, the collation unit 35 obtains the maximum value of the similarity and passes the identification information of the registered user corresponding to the maximum value together with the maximum value of the similarity to the authentication determination unit 36 of the processing unit 33.

The authentication determination unit 36 determines whether or not the maximum similarity value is equal to or greater than the authentication determination threshold value (step S211).
When the maximum value of the similarity is equal to or greater than the threshold for authentication determination (step S211—Yes), the authentication determination unit 36 authenticates the user as a registered user associated with the identification information corresponding to the maximum value of similarity. (Step S212).
On the other hand, when the maximum value of the similarity is less than the threshold value for authentication determination (step S211—No), the authentication determination unit 36 does not authenticate the user (step S213).
After step S212 or S213, the biometric authentication device 1 ends the biometric authentication process.
The biometric authentication device 1 may interchange the order of the process in step S203 and the processes in steps S204 to S206.

(registration process)
Next, a registration process for registering biometric information of a registered user as registered biometric information will be described.

FIG. 9 is an operation flowchart of the registration process.
The processing unit 24 of the terminal 2 acquires the user name of the user via an input unit (not shown) (Step S301). Further, the processing unit 24 acquires a biological image in which the user's biological information is captured from the biological information acquisition unit 21 (step S302). Then, the subject region detection unit 25 of the processing unit 24 detects a subject region in which a part including biological information is shown on the biological image (step S303). Thereafter, the collation data generation unit 26 of the processing unit 24 generates collation data for the biological information from the subject area (step S304).

  Further, the selection feature amount evaluation unit 27 of the processing unit 24 determines whether or not the selection feature amount can be extracted from the biological image according to the operation flowchart shown in FIG. 5 (step S305). When the selection feature quantity evaluation unit 27 determines that the biometric information needs to be reacquired, the processing unit 24 repeats the processing from step S302 onward.

On the other hand, when the selection feature quantity evaluation unit 27 determines that the selection feature quantity can be extracted, the selection feature quantity extraction unit 28 of the processing unit 24 obtains the selection feature quantity from the subject area (step S306).
Alternatively, if the selection feature quantity evaluation unit 27 determines that there is a possibility that the selection feature quantity may not be extracted, the alternative feature quantity extraction unit 29 of the processing unit 24 extracts the alternative feature quantity from the subject region and replaces it. Based on the feature value, an approximate value of the feature value for selection is obtained (step S307).

  After step S306 or S307, the processing unit 24 authenticates the verification data and the selection feature quantity or the approximate value of the selection feature quantity together with the user name of the user who wants to register via the communication network 5. 3 to send.

When the authentication server 3 receives the user verification data and the selection feature quantity or the approximate value of the selection feature quantity from the terminal 2, the registration unit 37 of the processing unit 33 is uniquely set for the user. Set the user identification number. Then, the registration unit 37 generates a registration request signal (step S308). The registration request signal includes identification information such as a user name of the user and a user identification number, verification data, and a selection feature amount. The registration unit 37 transmits the registration request signal to the data server 4 (step S309).
When receiving the registration request signal, the control unit 43 of the data server 4 stores various types of information, such as user identification information, collation data, and selection feature amounts, included in the registration request signal in the storage unit 42 (Step S43). S310). Then, the biometric authentication device 1 ends the registration process.
The biometric authentication device 1 may interchange the order of the process in step S303 and the processes in steps S304 to S307.

  As described above, the biometric authentication device according to this embodiment has a correlation with a selection feature amount when there is a possibility that a selection feature amount for selecting a registered user to be collated cannot be extracted. To extract high alternative features. The biometric authentication apparatus obtains an approximate value of the selection feature quantity based on the alternative feature quantity, and uses the approximate value for the selection of the registered user instead of the selection feature quantity. Even if it cannot be extracted, it is possible to appropriately select registered users to be collated.

  Next, a biometric authentication device according to the second embodiment will be described. The biometric authentication device according to the second embodiment extracts both the substitute feature quantity and the selection feature quantity when both the substitute feature quantity and the selection feature quantity can be extracted. The biometric authentication device uses the most probable feature quantity of the alternative feature quantity and the selection feature quantity in order to select a registered user to be collated. Below, among the components of the biometric authentication device according to the second embodiment, differences from the components of the biometric authentication device according to the first embodiment will be described.

FIG. 10 is a functional block diagram of the processing unit 241 of the terminal 2 according to the second embodiment. The processing unit 241 includes a subject area detection unit 25, a collation data generation unit 26, a selection feature quantity evaluation unit 27, a selection feature quantity extraction unit 28, an alternative feature quantity extraction unit 29, and a feature quantity selection unit. 30.
The processing unit 241 according to the second embodiment differs from the processing executed by the selection feature amount evaluation unit 27 in the feature amount selection unit 30 as compared with the processing unit 24 of the terminal according to the first embodiment. Therefore, hereinafter, the selection feature quantity evaluation unit 27 and the feature quantity selection unit 30 will be described. For the other components of the biometric authentication device, refer to the description of the corresponding components in the first embodiment.

  Even if there is a possibility that the selection feature quantity may not be extracted accurately, the selection feature quantity evaluation unit 27, if it is possible to extract the selection feature quantity, In addition to the extraction, the selection feature amount extraction unit 28 also extracts the selection feature amount. For example, if at least a part of the portion used for calculating the palm width is out of the allowable range in step S106 of the operation flowchart of the selection feature amount evaluation process, the selection feature amount evaluation unit 27 sets the alternative feature amount. As well as extracting feature quantities for selection. In this case, the selection feature quantity extraction unit 28 also stores the coordinates of the little finger side end point and the index finger side end point used in the calculation of the palm width together with the selection feature quantity in the storage unit 23. Similarly, the substitute feature amount extraction unit 29 is the coordinates of the position used for the substitute feature amount, for example, the distance between the midpoint of the base of the thumb and the index finger and the midpoint of the base of the ring finger and the little finger. In this case, the coordinates of the two middle points are stored in the storage unit 23 together with the approximate value of the selection feature amount.

  Also, in step S107 in the operation flowchart of the selection feature quantity evaluation process, even when the contour luminance difference is less than the predetermined luminance threshold, the selection feature quantity evaluation unit 27 not only selects the alternative feature quantity but also the selection feature quantity. It may be extracted. In this case, the selection feature amount extraction unit 28 calculates the average value of the luminance values of the pixels corresponding to the little finger side end points and the index finger side end points used for calculating the palm width and the background region adjacent to the pixels. The difference from the average value of the luminance values of the pixels is stored in the storage unit 23 together with the selection feature amount. Similarly, the alternative feature amount extraction unit 29 determines the difference between the average value of the luminance values of the pixels on the contour of the subject area used for the alternative feature amount and the average value of the luminance values of the pixels in the background area adjacent to those pixels. Are stored in the storage unit 23 together with the approximate value of the feature value for selection.

When both the selection feature quantity and the substitute feature quantity are calculated, the feature quantity selection unit 30 selects the more probable feature quantity of the two feature quantities, and selects the registered user to be collated. Is selected as a feature amount.
For example, when the coordinates of two pixels on the contour of the subject region used for calculation of the selection feature amount and the substitute feature amount are stored, the feature amount selection unit 30 stores the selection feature amount and the substitute feature amount. For each, the average value of the distance between these two pixels and the allowable range is calculated. When a pixel on the contour is included in the allowable range, the distance between the pixel and the allowable range is 0. The feature amount selection unit 30 selects the feature amount having the smaller average value of the distance between the two pixels on the contour of the subject region and the allowable range from the selection feature amount and the alternative feature amount as being more likely. Since the center of the biological image generally corresponds to a point located on the optical axis of the optical system of the biological information acquisition unit 21, it is not affected by the distortion of the optical system. Further, the center of the biological image is not affected by the decrease in the amount of peripheral light due to the optical system of the biological information acquisition unit 21. Therefore, the feature quantity selection unit 30 calculates the average value of the distance from the center of the biological image to the two pixels on the contour of the subject area used for calculating the feature quantity for selection, and the alternative feature quantity from the center of the biological image. The average value of the distances to two pixels on the contour of the subject area used for the calculation may be calculated. And the feature-value selection part 30 may select the feature-value with the smaller average value of the distance from the center of a biometric image as being more probable.

  Also, the difference between the average value of the luminance values of the pixels on the contour of the subject area used for the calculation of the selection feature amount and the alternative feature amount and the average value of the luminance values of the pixels in the adjacent background region is stored. If so, the feature value selection unit 30 selects the feature value having the larger difference between the average values of the luminance values.

  The processing unit 24 selects the selected feature amount of the selection feature amount and the alternative feature amount (however, if the alternative feature amount is selected, the approximate value of the selection feature amount) The data is transmitted to the authentication server 3 together with the verification data.

  According to the second embodiment, when the biometric authentication device can extract both the feature quantity for selection and the alternative feature quantity, the most probable one of these feature quantities can be used for the selection of the registered user to be collated. . Therefore, this biometric authentication device can more appropriately select registered users to be verified.

  Next, a biometric authentication device according to a third embodiment will be described. The biometric authentication apparatus according to the third embodiment uses a fingerprint together with a palm vein pattern or only a fingerprint for collation. This biometric authentication device uses the length of the finger with the vortex center of the fingerprint as one end as the selection feature quantity, and the finger with the centroid of the region where the fingerprint appears on the biometric image as one alternative feature quantity. Is used. Below, among the components of the biometric authentication device according to the third embodiment, differences from the components of the biometric authentication device according to the first embodiment will be described.

The biometric information acquisition unit 21 generates a second biometric image in which the fingerprint of any finger of the user is shown. For this purpose, the biological information acquisition unit 21 has a fingerprint sensor. The fingerprint sensor may be any fingerprint sensor that can take an image from the tip of any finger to the base of the finger, for example, and may be an optical path separation optical fingerprint sensor.
When not only the fingerprint but also the palm vein pattern is used for collation, the fingerprint sensor is arranged closer to the fingertip than the area sensor that captures the palm vein pattern. Then, the biometric information acquisition unit 21 generates a biometric image showing the palm vein pattern and a second biometric image showing the fingerprint.

The verification data generation unit 26 of the processing unit 24 generates verification data for a fingerprint from the second biological image.
For this purpose, the collation data generation unit 26 detects, for example, a pixel in which a ridge region and a valley line that are a set of pixels in which the ridge of the finger is reflected in the second biological image, and a user's finger. A background region that is a set of pixels that are not captured is distinguished.
For this purpose, the collation data generation unit 26 includes, for example, a pixel having a pixel value greater than or equal to the binarization threshold included in the ridge region and a pixel having a pixel value less than the binarization threshold included in the background region. The second biological image is binarized. The binarization threshold value is, for example, a predetermined fixed value (for example, 100), an average pixel value in the second biological image, or a pixel value included in the entire second biological image or a part thereof. It is set to a threshold value determined by analysis. Also, in the binarized second biological image, the value of the pixel included in the ridge area is set to “1”, for example, and the value of the pixel included in the background area is set to “0”, for example. .

  For example, when the biometric authentication device 1 performs matching processing by minutia matching, the collation data generation unit 26 extracts feature points such as ridge end points or branch points used in the minutia matching from the second biometric image. The collation data generation unit 26 uses the position of the feature point or the type of the feature point as collation data.

  For this purpose, the collation data generation unit 26 performs thinning processing on a set of pixels having pixel values corresponding to ridges on the binarized second biological image, and the ridges are thinned. A thinned binary image is generated. Then, the collation data generation unit 26 scans the thinned binary image using a plurality of templates corresponding to either the branch points or the end points of the ridges, thereby thinning when matching with any of the templates. A position on the binary image is detected. Then, the verification data generation unit 26 extracts the central pixel at the detected position as a minutiae. The template is represented by, for example, 3 × 3 pixels and has a binary pattern corresponding to a branch point or an end point of a ridge.

  The collation data generation unit 26 stores the type and coordinates of each feature point in the storage unit 23 as collation data.

The selection feature amount evaluation unit 27 determines whether or not the vortex center of the fingerprint on the second biological image can be extracted. If the vortex center cannot be extracted, the selection feature quantity evaluation unit 27 determines that the selection feature quantity cannot be extracted.
For example, the selection feature amount evaluating unit 27 detects the relative position between the template corresponding to the vortex center of the fingerprint and the thinned second biological image, as in the case of detecting the branch point or the end point of the ridge. While matching, pattern matching is performed. Thereby, the selection feature quantity evaluation unit 27 calculates a normalized cross-correlation value between the template and the thinned second biological image. If the maximum value of the normalized cross-correlation value is equal to or greater than a predetermined threshold (for example, 0.8), the selection feature amount evaluation unit 27 determines that the vortex center can be detected. Then, the selection feature quantity evaluation unit 27 causes the selection feature quantity extraction unit 28 to select the length of the finger with the vortex center as one end and the other end with a predetermined point closer to the base of the finger than the vortex center. Extract as a quantity. The predetermined point can be, for example, the first joint, the second joint, or the base of the finger.

  On the other hand, when the maximum value of the normalized cross-correlation value is less than the predetermined threshold, the selection feature value evaluation unit 27 determines that the vortex center cannot be detected. Then, the selection feature quantity evaluation unit 27 gives the substitute feature quantity extraction unit 29 the length of the finger with the centroid of the fingerprint as one end and the other end with a predetermined point on the finger base side from the centroid. Let's extract as

  Note that the selection feature amount evaluating unit 27 may determine whether or not the vortex center can be detected based on the quality of the second biological image, for example, the definition of the second biological image. In this case, for example, the selection feature quantity evaluation unit 27 performs a process of converting the second biological image from the spatial domain to the frequency domain, such as fast Fourier transform, and generates a specific frequency component corresponding to the fingerprint pattern. The value or the maximum value of the power spectrum is obtained as an evaluation value representing the sharpness.

Alternatively, the selection feature value evaluating unit 27 may calculate the difference between the local ridge direction and the smoothed ridge direction as the sharpness. In this case, for example, the selection feature amount evaluating unit 27 divides the thinned second biological image into a plurality of blocks, and calculates the ridge direction included in each block for each block. Each block has a size including a plurality of ridges, for example. Then, for each block, the selection feature value evaluating unit 27 sets an average value in a tangential direction of a line in which pixels representing ridges are continuous in the block as a ridge direction in the block.
Further, for each block, the selection feature value evaluating unit 27 sets the average value of the ridge directions of the block and the adjacent blocks as the smoothed ridge direction of the block. Then, the selection feature quantity evaluation unit 27 obtains an absolute value of a difference between the ridge direction obtained for each block and a smoothed ridge direction, and an absolute value of the difference is an evaluation value representing sharpness. And

  When the evaluation value is equal to or greater than a predetermined sharpness threshold value, the selection feature amount evaluation unit 27 extracts the length of the finger having the vortex center as one end as the selection feature amount in the selection feature amount extraction unit 28. Let On the other hand, if the evaluation value is less than the predetermined sharpness threshold value, the selection feature amount evaluation unit 27 instructs the alternative feature amount extraction unit 29 to use the length of the finger having the fingerprint centroid as one alternative feature amount. Let it be extracted.

The selection feature quantity extraction unit 28 uses the length from the vortex center of the fingerprint of a specific finger to a predetermined point closer to the base of the finger than the vortex center of the finger as the length of the finger that is the selection feature quantity. calculate.
For example, as described above, the selection feature amount extraction unit 28 performs the pattern matching between the template corresponding to the vortex center of the fingerprint and the thinned second biological image, and the thinned second biological image. The position that best matches the template above is detected as the vortex center. When the vortex center is detected by the selection feature quantity evaluation unit 27, the selection feature quantity extraction unit 28 may use the coordinates of the detected vortex center as they are.

  In addition, the selection feature amount extraction unit 28 divides, for example, the thinned second biological image into a plurality of small regions in order to detect the first joint of the finger. Each small region has a size that includes one or more ridges. Then, the selection feature amount extraction unit 28 obtains the ridge direction for each small region. In the vicinity of the first joint, there are many small regions in which the ridge direction is substantially orthogonal to the longitudinal direction of the finger, that is, the vertical direction of the second biological image. Therefore, the selection feature amount extraction unit 28 determines the number of small regions in which the ridge direction is substantially orthogonal to the vertical direction of the second biological image on the finger base side from the vortex center in each of the plurality of directions. For each line that is orthogonal to the direction and has the width of the small area. Then, the selection feature amount extraction unit 28 selects the line closest to the vortex center among the lines in which the number of small areas in which the ridge direction is substantially perpendicular to the vertical direction of the second biological image is the maximum value. Detect as one joint. Similarly, the selection feature quantity extraction unit 28 is the second from the vortex center among the lines in which the number of small regions in which the ridge direction is substantially orthogonal to the vertical direction of the second biological image is a maximum value. A close line may be detected as the second joint. Alternatively, the selection feature amount extraction unit 28 may detect the position of the first joint or the second joint according to another known method for detecting the position of the first joint or the second joint of the finger on the biological image. .

Further, the selection feature amount extraction unit 28 is the third closest to the center of the vortex among the lines in which the number of small regions in which the ridge direction is in a direction substantially orthogonal to the vertical direction of the second biological image is a maximum value. The line may be detected as the base of the finger. Alternatively, the selection feature amount extraction unit 28 obtains the width of the region in which the finger is captured for each row in order from the upper end of the second biological image, and further, the width of the region in the row and the row one level higher. Find the difference from the width of the region at. Then, the selection feature amount extraction unit 28 may have a root of a finger in a row where the difference is equal to or greater than a predetermined value.
Alternatively, when the biometric information acquisition unit 21 includes an area sensor that captures a palm vein pattern, the selection feature amount extraction unit 28 uses a finger on the biometric image generated by the area sensor. The root may be detected. Since the positional relationship between the area sensor and the fingerprint sensor is fixed, the selection feature amount extraction unit 28 can convert any pixel on the biological image into coordinates on the second biological image. The conversion formula is stored in the storage unit 23 in advance, for example. Then, the selection feature amount extraction unit 28 can calculate the length from the center of the vortex to the base of the finger by converting the position of the base of the finger on the biometric image into coordinates on the second biometric image.

The substitute feature amount extraction unit 29 calculates a length from the center of gravity of the fingerprint of a specific finger to a predetermined point closer to the base of the finger than the center of gravity of the finger as the length of the finger as the substitute feature amount.
For example, the alternative feature amount extraction unit 29 sets the center of gravity of the pixel corresponding to the ridge on the tip side of the finger relative to the first joint of the finger as the center of gravity of the fingerprint. The position of the first joint, the second joint, or the base of the finger can be detected by a method similar to the method described for the selection feature amount extraction unit 28. However, in this case, since the vortex center has not been detected, the alternative feature amount extraction unit 29 has the first joint, the second joint, or the finger below the position of the average distance from the tip of the finger to the vortex center. What is necessary is just to detect the root of.

Note that the center of gravity of the fingerprint is generally located near the vortex center of the fingerprint. Therefore, the alternative feature quantity extraction unit 29 sets the alternative feature quantity itself as an approximate value of the selection feature quantity.
As in the first embodiment, the processing unit 24 of the terminal 2 transmits the user verification data and the selection feature quantity or the approximate value of the selection feature quantity via the communication unit 22 when executing the authentication process. Transmit to the authentication server 3. Further, when executing the registration process, the processing unit 24 sends the user verification data and the selection feature quantity or the approximate value of the selection feature quantity together with the user name of the user to the authentication server 3 via the communication unit 22. Send.

  As with the selection unit 34 in the first embodiment, the selection unit 34 of the processing unit 33 of the authentication server 3 has, for each registered user, a feature amount for selecting the registered user and a feature amount for selecting the user. Alternatively, the absolute value of the difference from the approximate value is calculated as a selection score. And the selection part 34 selects the registration user used as collation object based on the score for selection.

  The collation unit 35 of the processing unit 33 is similar between the user biometric information and the selected biometric information of each registered user based on the user collation data and the selected registered user collation data. Each degree is calculated.

In the present embodiment, as in the first embodiment, when using the minutia matching, the collation unit 35 uses the fingerprint minutia included in the registered user collation data and the minutia included in the user collation data. Find the number of minutiae that match. The collation unit 35 can calculate the similarity between the user's fingerprint and the registered user's fingerprint by dividing the number of matching by the number of minutiae extracted for the user's fingerprint.
Alternatively, when using pattern matching, the matching unit 35 calculates the normalized cross-correlation value while changing the relative position between the region where the user's finger is captured and the region where the registered user's finger is captured. . Then, the matching unit 35 sets the maximum value of the normalized cross-correlation value as the similarity between the user's fingerprint and the registered user's fingerprint. Then, the collation unit 35 obtains the maximum value of the similarity, and passes the identification value of the registered user corresponding to the maximum value and the maximum value to the authentication determination unit 36.

  Alternatively, in the case where the palm vein pattern is also used for collation, the collation unit 35 calculates the palm vein pattern between the palm of the user and the palm vein pattern of the selected registered user as in the first embodiment. Similarity is also calculated. Then, the collation unit 35 obtains the maximum value of the sum of the fingerprint similarity and the palm vein pattern similarity, and passes the maximum value and the identification information of the registered user corresponding to the maximum value to the authentication determination unit 36.

  If the maximum value of the similarity or the maximum value of the similarity is equal to or greater than a predetermined authentication threshold, the authentication determination unit 36 corresponds the user to the maximum value of the similarity or the maximum value of the similarity. Authenticate as a registered user.

  According to this embodiment, when the biometric authentication device uses a fingerprint for biometric authentication, the selection feature amount can be obtained even if the selection feature amount for selecting the registered user to be collated is not accurately obtained. The registered user can be appropriately selected by using an alternative feature amount having a high correlation with.

  According to the modification of the third embodiment, the fingerprint sensor of the biometric information acquisition unit 21 has an area sensor that can simultaneously capture the fingerprints of a plurality of fingers, and the second living body in which the fingerprints of the plurality of fingers are reflected. It may be possible to generate an image. In this case, the selection feature amount evaluation unit 27 of the processing unit 24 of the terminal 2 uses the area from the first joint of the specific finger (hereinafter, referred to as an index as to whether or not the selection feature amount can be extracted). It may be used for convenience).

  For example, the selection feature amount evaluation unit 27 determines a set of pixels separated from each other by a pixel having a luminance value corresponding to the background from among a set of pixels having a luminance value corresponding to a ridge on the second biological image. Each finger region corresponds to one finger. Then, the selection feature amount evaluation unit 27 obtains the position of the first joint for each finger region, and calculates the number of pixels having a luminance value corresponding to the ridge ahead of the first joint as the fingerprint area of the finger. To do.

The selection feature quantity evaluation unit 27 determines that the selection feature quantity can be extracted when the fingerprint area of each finger photographed by the fingerprint sensor is equal to or greater than a predetermined area threshold. Then, the selection feature quantity evaluation unit 27 sets the length from the center of the vortex to the root of the finger or the second joint as the selection feature quantity for a specific finger. The area threshold value is set to, for example, the lower limit value of the fingerprint area necessary for all of the minutiae included in the fingerprint to appear in the second biological image.
On the other hand, if the fingerprint area for all the fingers is less than the area threshold value, the selection feature quantity evaluation unit 27 displays a message prompting the user to re-enter the fingerprint on the display unit of the terminal 2, and then The information acquisition unit 21 regenerates a second biological image obtained by photographing the user's finger.
In addition, regarding the finger whose selection feature quantity is to be calculated, if the fingerprint area is less than the area threshold value, but the fingerprint area of the other finger is equal to or larger than the area threshold value, the selection feature quantity evaluation unit 27 substitutes the alternative feature quantity. The extraction unit 29 is made to calculate a substitute feature amount.

  Alternatively, the selection feature amount evaluation unit 27 has an area ratio that is a ratio of the fingerprint area of the finger for obtaining the length as the selection feature amount with respect to the maximum value of the fingerprint areas of each finger on the second biological image. If it is equal to or greater than the predetermined ratio threshold, it is determined that the selection feature amount can be extracted. Then, the selection feature amount evaluation unit 27 may use the length from the center of the vortex to the base of the finger or the second joint as the selection feature amount. The ratio threshold value can be set to 0.5 to 0.8, for example. On the other hand, if the area ratio is less than the ratio threshold, the selection feature quantity evaluation unit 27 causes the substitution feature quantity extraction unit 29 to calculate a substitution feature quantity.

  In this case, the alternative feature quantity extraction unit 29 obtains the length from the first joint to the base of the finger or the second joint for the finger for which the selection feature quantity is obtained. Further, the alternative feature amount extraction unit 29 calculates the distance from the center of gravity of the fingerprint to the first joint for the finger having the maximum fingerprint area, and calculates the distance from the first joint of the finger for obtaining the selection feature amount. The value added to the base of the head or the length to the second joint is used as an alternative feature amount. Then, the alternative feature quantity extraction unit 29 sets the alternative feature quantity itself as an approximate value of the selection feature quantity.

  In the modification of each embodiment described above, the storage unit of the authentication server may store data for biometric authentication processing related to all registered users. In this case, the data server may be omitted. The terminal, the authentication server, and the data server may be integrated as one biometric authentication device. In this case, the processing unit of the integrated biometric authentication device realizes the function of the processing unit of the terminal and the function of the processing unit of the authentication server in each of the above embodiments. In addition, the storage unit of the integrated biometric authentication device stores data for biometric authentication processing related to all registered users.

  According to yet another modification, a registration device for registering user verification data and selection feature quantities in the data server can be connected to the data server separately from the terminal and the authentication server. Good. In this case, the registration device has each component of the terminal, and the processing unit of the registration device realizes the function of the registration unit of the authentication server.

  Furthermore, the computer program for causing the computer to realize each function of the processing unit of the terminal and each function of the processing unit of the authentication server is recorded in a computer-readable recording medium such as a magnetic recording medium or an optical recording medium. May be provided. However, the recording medium does not include a carrier wave.

All examples and specific terms listed herein are intended for instructional purposes to help the reader understand the concepts contributed by the inventor to the present invention and the promotion of the technology. It should be construed that it is not limited to the construction of any example herein, such specific examples and conditions, with respect to showing the superiority and inferiority of the present invention. Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions and modifications can be made thereto without departing from the spirit and scope of the present invention.
The following supplementary notes are further disclosed regarding the embodiment described above and its modifications.
(Appendix 1)
A biological information acquisition unit that generates a biological image in which a part including the biological information of the user is captured;
A collation data generation unit that generates collation data representing characteristics of the biometric information from the biometric image;
A feature extraction unit for selection that extracts a feature for selection representing the feature of the shape of the part from the biological image;
By extracting a substitute feature amount representing a feature of another shape of the part having a predetermined correlation with the selection feature amount from the biological image, and correcting the substitute feature amount according to the correlation An alternative feature quantity extraction unit that calculates an approximate value of the feature quantity for selection;
It is determined whether or not the selection feature amount can be extracted based on the quality of the part in the biological image, and when it is determined that the selection feature amount can be extracted, the selection feature amount extraction unit determines the selection feature amount. A feature quantity extraction unit, on the other hand, if it is determined that the selection feature quantity cannot be extracted, a selection feature quantity evaluation unit that causes the alternative feature quantity extraction unit to calculate an approximate value of the selection feature quantity;
For each of a plurality of registered users, a storage unit that stores the matching user data and the selection feature amount of the registered user;
A selection unit that selects a registered user having the selection feature amount similar to the selection feature amount of the user or the approximate value of the selection feature amount among the plurality of registered users;
A collation unit that collates the collation data of the selected registered user with the collation data of the user;
A biometric authentication device.
(Appendix 2)
The selection feature quantity evaluation unit causes the selection feature quantity extraction unit to extract the selection feature quantity when both the selection feature quantity and the substitution feature quantity can be extracted, and causes the substitution feature quantity extraction unit to Calculating an approximate value of the feature for selection;
The biometric authentication device according to supplementary note 1, further comprising a feature amount selection unit that selects a probable one of the selection feature amount and an approximate value of the selection feature amount for selection by the registered user.
(Appendix 3)
The selection feature quantity evaluation unit causes the alternative feature quantity extraction unit to calculate an approximate value of the selection feature quantity when a part of the part related to the selection feature quantity cannot be detected from the biological image. The biometric authentication device according to appendix 1 or 2.
(Appendix 4)
The part including the biological information is a hand, the selection feature quantity extraction unit obtains a palm width on the biological image as the selection feature quantity, and the substitute feature quantity extraction unit uses the substitute feature quantity as the substitution feature quantity. The width of any finger on the biometric image, or the distance between the midpoint of the base of the index finger and thumb and the midpoint of the base of the ring finger and little finger, or the user places his / her hand on the biometric information acquisition unit The biometric authentication device according to any one of appendices 1 to 3, wherein an interval between a left guide member positioned at a left end of the biometric information acquisition unit and a right guide member positioned at a right end of the biometric information acquisition unit is obtained. .
(Appendix 5)
The selection feature quantity evaluation unit extracts the substitute feature quantity when at least a part of the user's hand is out of an allowable range on the biological image in which distortion of the shape of the object shown in the biological image can be ignored. 5. The information processing apparatus according to claim 4, further comprising: determining a width of any finger on the biological image or a distance between a midpoint of the base of the index finger and the thumb and a midpoint of the base of the ring finger and the little finger as the substitute feature amount. Biometric authentication device.
(Appendix 6)
The biological information acquisition unit includes an area sensor and an optical system that forms an image of the part on the area sensor, and the selection feature amount evaluation unit is configured to reduce a light amount due to the optical system. When at least a part of the user's hand is out of the second allowable range on the biological image that does not affect the detection of the contour of the hand of the user, the biological feature is extracted as the alternative feature amount in the alternative feature amount extraction unit. The biometric authentication device according to appendix 4, wherein the biometric authentication apparatus is configured to obtain a width of any finger on the image or a distance between a midpoint of the base of the index finger and the thumb and a midpoint of the base of the ring finger and the little finger.
(Appendix 7)
When the alternative feature amount extraction unit obtains the width of any finger as the alternative feature amount, the feature for selection is obtained by multiplying the width of the finger by the ratio of the width of the palm to the width of one finger. The biometric authentication device according to any one of appendices 4 to 6, which calculates an approximate value of the quantity.
(Appendix 8)
When the alternative feature amount extraction unit obtains the distance between the midpoint of the base of the index finger and thumb and the midpoint of the base of the ring finger and little finger as the alternative feature amount, the ratio of the palm width to the distance is calculated as the distance. The biometric authentication device according to any one of appendices 4 to 6, wherein an approximate value of the feature value for selection is calculated by multiplying.
(Appendix 9)
The selection feature quantity evaluation unit has a left end of a region where a user's hand is shown on the biological image coincides with a position of the left guide member on the biological image, and the user on the biological image. When the right end of the region where the hand is shown coincides with the position of the right guide member on the biological image, the substitute feature amount extraction unit is provided with the space between the right guide member and the left guide member on the biological image. The biometric authentication device according to appendix 4, wherein the interval is obtained as an approximate value of the selection feature amount.
(Appendix 10)
The part including the biological information is a finger, and the selection feature amount extraction unit obtains the length of the finger with one end of the vortex center of the fingerprint on the biological image as the selection feature amount, and the substitute feature amount The biometric authentication device according to any one of appendices 1 to 3, wherein the extraction unit obtains a finger length having one end at a center of gravity of the fingerprint region on the biometric image as an approximate value of the feature value for selection.
(Appendix 11)
If the selection feature quantity evaluation unit cannot detect the vortex center of the fingerprint, the alternative feature quantity extraction unit sets the length of the finger with the center of the fingerprint area on the biological image as one end as the substitute feature quantity. The biometric authentication device according to appendix 10, which is requested.
(Appendix 12)
The biometric information acquisition unit generates the biometric image showing a plurality of fingers of the user,
If the area of the first finger among the plurality of fingers of the user on the biometric image is so small that the feature of the fingerprint of the finger cannot be extracted, the selection feature amount evaluation unit To calculate an approximate value of the feature for selection,
The alternative feature amount extraction unit has a length from the first joint of the first finger to a predetermined point closer to the root side than the first joint of the first finger. The biometric authentication device according to appendix 10, wherein a value obtained by adding a length from the center of gravity of the fingerprint of the finger to the first joint of the second finger is an approximate value of the selection feature amount.
(Appendix 13)
A terminal,
An authentication server connected to the terminal via a communication network;
The terminal includes the collation data generation unit, the selection feature amount evaluation unit, the selection feature amount extraction unit, and the alternative feature amount extraction unit.
The biometric authentication device according to any one of appendices 1 to 12, wherein the authentication server includes the selection unit and the verification unit.
(Appendix 14)
A biological information acquisition unit that generates a biological image in which a part including the biological information of the user is captured;
A collation data generation unit that generates collation data representing characteristics of the biometric information from the biometric image;
A feature extraction unit for selection that extracts a feature for selection representing the feature of the shape of the part from the biological image;
By extracting a substitute feature amount representing a feature of another shape of the part having a predetermined correlation with the selection feature amount from the biological image, and correcting the substitute feature amount according to the correlation An alternative feature quantity extraction unit that calculates an approximate value of the feature quantity for selection;
It is determined whether or not the selection feature amount can be extracted based on the quality of the part in the biological image, and when it is determined that the selection feature amount can be extracted, the selection feature amount extraction unit determines the selection feature amount. A feature quantity extraction unit, on the other hand, if it is determined that the selection feature quantity cannot be extracted, a selection feature quantity evaluation unit that causes the alternative feature quantity extraction unit to calculate an approximate value of the selection feature quantity;
A registration unit that associates the identification information of the user with the matching data and the selection feature quantity or an approximate value of the selection feature quantity in a storage unit;
A biometric information registration device.
(Appendix 15)
Generate a biological image showing the part containing the user's biological information,
From the biometric image, generate verification data representing the characteristics of the biometric information,
Based on the quality of the part in the biological image, it is determined whether or not a feature for selection representing the feature of the shape of the part can be extracted. On the other hand, if it is determined that the selection feature quantity cannot be extracted, an alternative feature quantity representing a feature of another shape of the part having a predetermined correlation with the selection feature quantity is extracted. And calculating an approximate value of the feature for selection by correcting the substitute feature according to the correlation,
Of the plurality of registered users, select a registered user having the selection feature amount similar to the selection feature amount of the user or an approximate value of the selection feature amount;
Collating the collation data of the selected registered user with the collation data of the user;
A biometric authentication method.
(Appendix 16)
From the biometric image in which the part including the biometric information of the user is captured, the verification data representing the characteristics of the biometric information is generated,
Based on the quality of the part in the biological image, it is determined whether or not a feature for selection representing the feature of the shape of the part can be extracted. On the other hand, if it is determined that the selection feature quantity cannot be extracted, an alternative feature quantity representing a feature of another shape of the part having a predetermined correlation with the selection feature quantity is extracted. And calculating an approximate value of the feature for selection by correcting the substitute feature according to the correlation,
Of the plurality of registered users, select a registered user having the selection feature amount similar to the selection feature amount of the user or an approximate value of the selection feature amount;
Collating the collation data of the selected registered user with the collation data of the user;
A computer program for biometric authentication for causing a computer to execute this.

DESCRIPTION OF SYMBOLS 1 Biometric authentication apparatus 2 Terminal 21 Biometric information acquisition part 22 Communication part 23 Storage part 24 Processing part 25 Subject area detection part 26 Collation data generation part 27 Selection feature-value evaluation part 28 Selection feature-value extraction part 29 Alternative feature-value extraction Unit 30 feature amount selection unit 3 authentication server 31 communication unit 32 storage unit 33 processing unit 34 selection unit 35 verification unit 36 authentication determination unit 37 registration unit 4 data server 41 communication unit 42 storage unit 43 control unit 5 communication network

Claims (7)

A biological information acquisition unit that generates a biological image in which a part including the biological information of the user is captured;
A collation data generation unit that generates collation data representing characteristics of the biometric information from the biometric image;
A feature extraction unit for selection that extracts a feature for selection representing the feature of the shape of the part from the biological image;
By extracting a substitute feature amount representing a feature of another shape of the part having a predetermined correlation with the selection feature amount from the biological image, and correcting the substitute feature amount according to the correlation An alternative feature quantity extraction unit that calculates an approximate value of the feature quantity for selection;
It is determined whether or not the selection feature amount can be extracted based on the quality of the part in the biological image, and when it is determined that the selection feature amount can be extracted, the selection feature amount extraction unit determines the selection feature amount. A feature quantity extraction unit, on the other hand, if it is determined that the selection feature quantity cannot be extracted, a selection feature quantity evaluation unit that causes the alternative feature quantity extraction unit to calculate an approximate value of the selection feature quantity;
For each of a plurality of registered users, a storage unit that stores the matching user data and the selection feature amount of the registered user;
A selection unit that selects a registered user having the selection feature amount similar to the selection feature amount of the user or the approximate value of the selection feature amount among the plurality of registered users;
A collation unit that collates the collation data of the selected registered user with the collation data of the user;
A biometric authentication device. The selection feature quantity evaluation unit causes the selection feature quantity extraction unit to extract the selection feature quantity when both the selection feature quantity and the substitution feature quantity can be extracted, and causes the substitution feature quantity extraction unit to Calculating an approximate value of the feature for selection;
The biometric authentication device according to claim 1, further comprising a feature quantity selection unit that selects a probable one of the selection feature quantity and an approximate value of the selection feature quantity to be used for selection by the registered user.   The selection feature quantity evaluation unit causes the alternative feature quantity extraction unit to calculate an approximate value of the selection feature quantity when a part of the part related to the selection feature quantity cannot be detected from the biological image. The biometric authentication device according to claim 1.   The part including the biological information is a hand, the selection feature quantity extraction unit obtains a palm width on the biological image as the selection feature quantity, and the substitute feature quantity extraction unit uses the substitute feature quantity as the substitution feature quantity. The width of any finger on the biometric image, or the distance between the midpoint of the base of the index finger and thumb and the midpoint of the base of the ring finger and little finger, or the user places his / her hand on the biometric information acquisition unit The biometric authentication according to any one of claims 1 to 3, wherein an interval between a left guide member positioned at a left end of the biometric information acquisition unit and a right guide member positioned at a right end of the biometric information acquisition unit is obtained. apparatus.   The part including the biological information is a finger, and the selection feature amount extraction unit obtains the length of the finger with one end of the vortex center of the fingerprint on the biological image as the selection feature amount, and the substitute feature amount The biometric authentication device according to any one of claims 1 to 3, wherein the extraction unit obtains a finger length having one end at a center of gravity of the fingerprint region on the biometric image as an approximate value of the selection feature amount. Generate a biological image showing the part containing the user's biological information,
From the biometric image, generate verification data representing the characteristics of the biometric information,
Based on the quality of the part in the biological image, it is determined whether or not a feature for selection representing the feature of the shape of the part can be extracted, and when it is determined that the feature for selection can be extracted, the selection feature On the other hand, if it is determined that the selection feature quantity cannot be extracted, an alternative feature quantity representing a feature of another shape of the part having a predetermined correlation with the selection feature quantity is extracted. And calculating an approximate value of the feature for selection by correcting the substitute feature according to the correlation ,
Of multiple registered users, select the registered user with the selected feature quantity that is similar to the approximation of the selected feature quantity or the selected feature quantity of the user,
Collating the collation data of the selected registered user with the collation data of the user;
A biometric authentication method. From the biometric image in which the part including the biometric information of the user is captured, the verification data representing the characteristics of the biometric information is generated
Based on the quality of the part in the biological image, it is determined whether or not a feature for selection representing the feature of the shape of the part can be extracted, and when it is determined that the feature for selection can be extracted, the selection feature On the other hand, if it is determined that the selection feature quantity cannot be extracted, an alternative feature quantity representing a feature of another shape of the part having a predetermined correlation with the selection feature quantity is extracted. And calculating an approximate value of the feature for selection by correcting the substitute feature according to the correlation ,
Of multiple registered users, select the registered user with the selected feature quantity that is similar to the approximation of the selected feature quantity or the selected feature quantity of the user,
Collating the collation data of the selected registered user with the collation data of the user;
A computer program for biometric authentication for causing a computer to execute this.