JP2007052534A - Palm print authentication device, palm print authentication program, palm print authentication method, palm print image extraction method, and portable telephone terminal with palm print authentication device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a palm print authentication device capable of performing an authentication by using the palm print of a portion of the palm of the hand collected by an image reading sensor whose size is fitted to the palm of the hand, and to provide a palm print authentication program, a palm print authentication method, a palm print image extraction method and a portable telephone terminal with the palm print authentication device. <P>SOLUTION: An image recognizing part detects a valley section between the middle finger and the third finger from the partial image of the palm of the hand of a user adopted by a palm print sensor as a reference point P, and an image selecting part selects a square TUVW whose one side is a segment passing the reference point P, which is parallel with and as long as the short side of the image, as a section to be used for collation. Then, the palm print is extracted by detecting the edge of the image in the selected square TUVW, and collated with master data. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to palm print authentication that performs electronic authentication using a palm palm print, and more particularly to palm print authentication that performs authentication using a palm print of a part of the palm.

In recent years, electronic personal authentication has been widely used in various daily situations such as electronic commerce and entrance / exit management. For such personal authentication, a simple method using an ID card or a password has been used for some time, but fingerprints, irises of eyes, voices, etc. are methods that can ensure security with higher reliability. So-called biometric authentication that confirms the identity of a person by using human physical characteristics has also been realized.
Palm prints of palms, that is, patterns composed of “main lines” such as emotion lines and life lines, other fine “wrinkles”, and “ridge lines” on the skin, can be used as biometric information for identifying individuals in the same way as fingerprints. In addition, unlike the fingertip fingerprint, it has a feature that there is little “fading” due to aging and wear. For this reason, palmprint authentication technology has been proposed as one of biometric authentication (see Non-Patent Document 1, for example).

  Here, a palmprint authentication method proposed in Non-Patent Document 1 will be described. This authentication method consists of four processes: collection of a palm image, setting of an image area used for authentication, conversion from image data to a bit matrix, and collation between a bit matrix and registered data. We will explain in the following order. In the present specification, the term “palmprint” is defined as a word including the “main line”, “wrinkle”, and “lift line”.

(1) Image collection In the method of Non-Patent Document 1, image data of the entire palm is collected using a dedicated scanning device constituted by a CCD camera or the like. An example of the collected image is shown in FIG.

(2) Authentication area setting In Non-Patent Document 1, instead of using the entire collected palm image, a partial image near the center of the palm where “main lines” and “wrinkles” are gathered is used. Authentication is performed. This image area is an area indicated by a broken line in FIG. 5 and is determined by the following processing.
First, a valley portion (referred to as point A) between the base of the index finger and the middle finger and a valley portion (referred to as point B) between the base of the ring finger and the little finger are extracted from the image of the entire palm. This is processing by a known image recognition technique. When a straight line connecting these two points A and B is taken as the y-axis, and a straight line passing through the middle point between the two points A and B is taken as the x-axis, the center is on the x-axis. Set a square whose sides are parallel to the x-axis. This square is the area of the broken line. However, the size of the square (the length of one side) is a parameter determined by an empirical value. The image data of the authentication square area set in this way is cut out from the entire image of the palm of FIG.

(3) Conversion to Bit Matrix The extracted image data for authentication is subjected to image convolution processing by using a Gabor filter. The Gabor filter is a spatial filter often used to extract an edge having a specific direction and width (period) from a given image, and a two-dimensional Gabor function is expressed by the following equation.

  In the above equation, i is an imaginary unit, u is the spatial frequency, θ is the direction of the edge to be extracted, and σ is the half-value width of the Gaussian function. If the image data for authentication is expressed as F (x, y), the convolution H (x, y) with the Gabor function G can be written as the following equation. However, the Gabor function may be applied to the following (Formula 2) after being appropriately corrected (coordinate system conversion, etc.).

  Here, the sum about p and q shall be performed with respect to all the pixels of the image data for authentication. By the above calculation, edges (such as “main line” and “wrinkle”) included in the image are extracted from the authentication image (the palm print in the square indicated by the broken line in FIG. 5). FIG. 6 is an example of an image after extraction. As a result of the edge extraction, the image is expressed as a bit matrix consisting of binary values of white and black.

(4) Data collation The bit matrix obtained as described above is collated with master data that has been processed and registered in the same procedure in advance. The collation is performed by calculating a normalized Hamming distance between the bit matrix (P) to be authenticated and the bit matrix (Q) of the master data. Here, the Hamming distance is obtained by comparing the corresponding bits of the two bit matrices to obtain the number of different bits. The normalized Hamming distance D between P and Q is expressed by the following equation.

In the above equation, the subscripts R and I represent the real part and the imaginary part, respectively, and N is the number of pixels in the vertical (or horizontal) direction. As is clear from the definition, the normalized Hamming distance D becomes smaller as the similarity between the bit matrices P and Q becomes higher (D = 0 when they are completely the same). Therefore, a threshold D ₀ is provided, and D ≦ In the case of D ₀ , authentication is performed assuming that the authentication image matches the master data. D> In the case of _{D 0} authentication is not performed.
David Zhang et al., “Online Palmprint Identification”, IEEE Transactions on Pattern Analysis and Machine Intelligence, September 2003, Vol. 25, No. 9, pp. 1041-1050

As described above, palmprint authentication can be realized using the method described in Non-Patent Document 1. However, in this method, since an image of the entire palm is collected, the image reading unit of the scanning device needs to be several tens of centimeters square, and there is a problem that the system cannot be reduced in size. Further, since it is necessary to correctly align the palm palm portion of the palm with the image reading unit at the time of image collection, there is a demerit that a person who is authenticated must perform an authentication operation actively and consciously.
Also, if the scanning device is downsized, an image of the entire palm cannot be obtained in this case, so that the image for authentication cannot be extracted by the above method (2), and palmprint authentication is performed. It becomes impossible to do.

  The present invention has been made in view of these problems, and an object of the present invention is to perform palmprint authentication that can be performed by using a palmprint of a part of the palm collected by an image reading sensor having a size that fits in the palm of the hand. An apparatus, a palm print authentication program, and a palm print authentication method are provided. Also provided are a palmprint image extraction method for extracting the palmprint and a mobile phone terminal provided with a palmprint authentication apparatus.

  The present invention has been made to solve the above-mentioned problems, and the invention according to claim 1 is a sensor for collecting an image including a palm print from the palm, and between the middle finger and the base of the ring finger from the collected image. An image recognition means for detecting a valley portion of the image as a reference point, and a square image having a line segment passing through the reference point and parallel to the short side of the image and having the same length as one side in the image. An image selecting means for selecting the image, a palm pattern in the selected image is extracted by detecting an edge, and the image converting means for converting the image into an image composed of edges, and a collation for a specific registered person Storage means for storing an image for use, and checking means for checking the converted image with the checking image and performing authentication when the two images are considered to match based on a predetermined threshold. Palm characterized by It is an authentication device.

  According to a second aspect of the present invention, the computer detects a valley portion between the middle finger and the base of the ring finger from the input image including the palm print as a reference point, and a line passing through the reference point in the image. Selecting a square image having a side that is a line segment of the same length and parallel to the short side of the image, and extracting a palm print in the selected image by detecting an edge; Is converted into an image composed of edges, the converted image is compared with an image for verification relating to a specific registered person stored in the storage unit, and, as a result of the verification, predetermined And a step of performing authentication when it is considered that the images match with each other as a reference.

  According to a third aspect of the present invention, there is provided a step of collecting an image including a palm print from a palm using a sensor, and image recognition by using a valley portion between a middle finger and a base of a ring finger from the collected image as a reference point. Selecting a square image having a line segment passing through the reference point and parallel to the short side of the image and having the same length as one side in the image; and Extracting the palm print in the image, collating the extracted palm print with a palm print relating to a specific registered person stored in the storage unit, and as a result of the collation, a predetermined threshold is used as a reference. And a step of performing authentication when it is considered that the palm prints coincide with each other.

  According to a fourth aspect of the present invention, a step of detecting a valley portion between the middle finger and the base of the ring finger from an image including a palm print by image recognition as a reference point, and a line passing through the reference point in the image And a step of extracting a square image having one side of a line segment that is parallel to the short side of the image and having the same length as one side.

  The invention according to claim 5 is a mobile phone terminal comprising the palm print authentication device according to claim 1, wherein the sensor in the palm print authentication device is a housing provided with a main operation unit such as a numeric keypad. The mobile phone terminal is disposed on a surface opposite to the body surface.

  According to the present invention, in a palm image collected using a sensor, a valley portion between the middle finger and the base of the ring finger is detected as a reference point, a line segment passing through the reference point, and the short side of the image Since a square image with a parallel line segment of the same length as one side is selected and palm print authentication is performed using the selected image, the palm print required for authentication can be obtained from an image of a part of the palm. Information (the square image) can be accurately extracted. Since the image to be collected only needs to include a part near the center of the palm and the reference point, the sensor can be downsized to a size that fits in the palm. In addition, since the sensor can be made small, a palmprint authentication function can be installed in a small information terminal such as a mobile phone, and security is ensured even if the terminal is lost or stolen. The

  Furthermore, according to the invention described in claim 5, since the sensor is arranged at the part that comes in contact with the palm of the hand when the terminal is held by hand, the user does not consciously perform the authentication operation but only holds the terminal. It is possible to perform authentication processing passively and automatically, or to receive authentication while performing other operations (for example, call operation).

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram showing a configuration of a palmprint authentication apparatus 1 according to an embodiment of the present invention. The palm print authentication device 1 is a general mobile phone terminal having a palm print authentication function, and includes a telephone function unit 10, a palm print sensor 21, an image recognition unit 22, an image selection unit 23, and an image. The conversion part 24, the collation part 25, and the palm print DB (database) 26 are comprised. In addition, this palmprint authentication apparatus 1 has the same shape as a general foldable mobile phone terminal in its outer shape (see FIG. 2).

  The telephone function unit 10 is a part that exhibits a function as a general mobile phone, and is a microphone that captures sound and outputs it as an electrical signal, a speaker that produces an input signal as sound, an input signal from the microphone and a speaker Audio processing circuit that performs audio processing of output signal, modulation / demodulation circuit and radio circuit that performs modulation processing of transmission signal and demodulation processing of received signal, operation unit consisting of antenna for transmitting and receiving radio waves, numeric keypad, etc., liquid crystal that displays various information It consists of a display, other operation control CPUs, cameras, vibrators, and the like. Note that each of the above components constituting the telephone function unit 10 is a general component of the mobile phone terminal and is not related to the essence of the present invention. Not done.

  The palm print sensor 21 is an optical sensor that collects a palm palm print as image data. As shown in FIG. 2, the palm print sensor 21 is disposed on the back of the casing provided with a numeric keypad in the palm print authentication apparatus 1. The size is about two-thirds of the back of the casing. Since it is designed in such an arrangement and size, the palmprint sensor 21 is placed in the palm of the hand when the mobile phone terminal (palmprint authentication device 1) is held in the right hand so that the user can operate the numeric keypad. On the other hand, it is in contact with a rectangular portion (the same shape as the sensor) indicated by a broken line in FIG. In other words, when the hand is held in such a manner, a palm print image of a rectangular area portion inclined by approximately 45 degrees with respect to the palm axis direction (the direction in which the little finger grows from the index finger) is collected by the palm print sensor 21. It has become.

  The image reading by the palmprint sensor 21 is performed by emitting laser light from a laser diode in the sensor, scanning the object, measuring reflected light from the object, and identifying the magnitude of the light intensity. Is going.

Returning to FIG. 1, the description will be continued.
The image recognizing unit 22 detects a point P serving as a reference for extracting a partial image actually used for collation of the palm print from the palm palm print image (broken rectangle image of FIG. 2) collected by the palm print sensor 21. Perform image processing. As the reference point, the palm print authentication apparatus 1 uses a valley portion between the middle finger and the base of the ring finger. FIG. 3 shows an example of the collected palmprint image and the reference point P detected in the image.

  For the detection of the reference point P, a general image recognition method which is a well-known conventional technique is used. For example, a contour line (edge) is extracted from a palm print image and a portion having a curved shape of a valley between the fingers is detected, or a large number of palm print images (portions corresponding to FIG. 3) are previously used as reference images. For example, a method of detecting a reference point by performing pattern matching between the collected palmprint image and the reference image can be applied.

  The image selection unit 23 performs a process of determining a part that is actually used for collation of the palm print from the palm print image of FIG. Specifically, first, a straight line TU passing through the reference point P detected by the image recognition unit 22 and parallel to one side JK of the palm print image of FIG. 3 is a midpoint of the line segment TU perpendicular to the y axis and the y axis. An orthogonal coordinate system is set with the straight line passing through x as the x-axis. Then, an image in the square TUWW with one side of the line segment TU is selected as an image that is actually used for collation of palm prints.

  Here, in the square TUWW selected by the image selection unit 23 by such an algorithm, as can be seen from FIG. 3, many representative “main lines” and “wrinkles” are gathered. That is, by selecting a collation area (square TUWW) using this algorithm, it is necessary to perform collation of palm prints, which will be described later, with high accuracy from an image collected using a palm print sensor 21 that is at most the size of a mobile phone terminal. It is possible to acquire a sufficient amount of palm print information.

  The image conversion unit 24 performs an operation of extracting an edge in the image by applying a spatial filter function to the data of the square TUWW image selected by the image selection unit 23. As the spatial filter function, the two-dimensional Gabor function expressed by the above-described (Equation 1) is adopted as in the conventional technique described in Non-Patent Document 1. A specific equation for edge extraction uses the same (Formula 2) as in the prior art. That is, the convolution H (x, y) of the image data F (x, y) of the square area TUUV by the Gabor function G (x, y, θ, u, σ) in the xy coordinate system described above is the image conversion unit 24. By this calculation, the palm prints included in the image are extracted as edges.

  The palmprint image extracted as an edge is, for example, as shown in FIG. 6 (however, this figure is an example and does not correspond to the palmprint of FIG. 3). As shown in the extraction result of FIG. 6, a palm print converted into a black and white image composed of N pixels vertically and horizontally is obtained. Here, N is a design parameter selected so that the authentication described later succeeds with a certain accuracy. The N × N pixel data obtained in this way is used in subsequent processing as a bit matrix of N rows and N columns having 0 and 1 elements corresponding to black and white.

  The palm print DB 26 is a storage device that stores, as master data, palm print data processed by the palm print sensor 21, image recognition unit 22, image selection unit 23, and image conversion unit 24 for a specific person. It is called and used during the palmprint matching process performed in the matching unit 25. In addition, since the palmprint authentication apparatus 1 of this embodiment is a mobile phone terminal, it is assumed that the number of persons who register master data is 1 (user of mobile phone). However, there may be a plurality of registered persons. In this case, since the master data of the plurality of registered persons is used for verification, they are sequentially called.

The collation unit 25 performs an operation of collating the master matrix stored in the palm print DB 26 on the bit matrix input from the image conversion unit 24, and applies a predetermined calculation result to the bit matrix. Assuming that the palm print matches the registered master data, the person is authenticated as the registrant himself. Specifically, the normalized Hamming distance D of PQ is calculated using the same method as in Non-Patent Document 1, that is, the bit matrix to be authenticated is P and the bit matrix of the master data is Q, using the above-described (Equation 3). Is done. Then, authentication is given to the person only when D ≦ D ₀ with respect to the predetermined threshold value D ₀ .

Next, the operation of the palm print authentication apparatus 1 described above will be described using the flowchart shown in FIG.
First, a registration mode for registering a user of a mobile phone terminal (palmprint authentication device 1) as an initial setting will be described (FIG. 4A). First, when the user inputs a predetermined command using the numeric keypad or the like of the operation unit, the palm print authentication device 1 is set to the registration mode (step S10). Here, it is assumed that the predetermined command is known only to a user having a proper authority. That is, other than the regular user cannot change the mode of the palm print authentication device 1 to the registration mode.

In the registration mode, the palm print sensor 21 acquires a palm print image of the user's palm (step S12). As described above, the acquired image is a broken line portion in FIG.
Then, the image is sequentially processed by the image recognition unit 22, the image selection unit 23, and the image conversion unit 24 (step S13 to step S15), and a bit matrix of N rows and N columns is generated. The generated bit matrix is stored as master data in the palm print DB 26 (step S16), and the registration mode ends. As a result, the palm print of the user is registered in the palm print authentication apparatus 1, and thereafter, the apparatus 1 shifts to a state in which no operation is possible, i.e., the authentication mode unless the authentication procedure described below is performed.

  Next, an authentication process executed when the mobile phone terminal (palmprint authentication device 1) is actually used in the authentication mode will be described (FIG. 4B). First, when a user (which may be a regular user who has registered a palm print or another person) holds the palm print authentication apparatus 1 in his hand and presses an arbitrary key (step S11), the palm print sensor 21 detects the palm of the user's palm. A palm print image is acquired (step S12).

  The acquired image is sequentially processed by the image recognition unit 22, the image selection unit 23, and the image conversion unit 24 in the same manner as in the registration mode (step S13 to step S15), and a bit matrix of N rows and N columns is generated.

  Then, the collation unit 25 collates the generated bit matrix with the bit matrix of the master data called from the palm print DB 26 (step S17), and when it is considered that both match (in step S18, “ Yes "), authentication is established, and the mobile phone terminal becomes operable (step S19). On the other hand, if it is not considered that the two match (“No” in step S18), the authentication is rejected and the mobile phone terminal is maintained in an inoperable state (step S20).

  As described above, according to the above-described embodiment, the palm print sensor 21 captures the palm image of the user, and the image recognition unit 22 detects the valley portion between the middle finger and the base of the ring finger as the reference point P from the image. Then, in the image selection unit 23, a square TUWW passing through the reference point P and parallel to the short side of the image and having a line segment having the same length as one side is selected as a portion used for collation. As a result, from the image of a part of the palm of the minimum necessary size including the reference point P and the center of the palm, a portion of the palm print having a lot of “main line” and “wrinkle” information is extracted as a matching image. It is possible. According to the above embodiment, the palm print is extracted by detecting the edge of the image in the selected square TUUV, and collation with the master data is performed. Therefore, since it is sufficient for the sensor to be able to collect an image of a part of the palm, a small sensor can be used, and the palm print authentication device 1 can be mounted on a small electronic device such as a mobile phone terminal.

  In addition, when a user has a mobile phone terminal, since the way of holding the side of the terminal against the bulge near the base of the thumb is natural, the positional relationship between the palm and the mobile phone terminal (palmprint authentication device 1) The images collected are always in the same range, and there is an advantage that the accuracy of authentication is improved.

As described above, the embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to the above, and various design changes and the like can be made without departing from the scope of the present invention. It is possible to
For example, the palm print authentication apparatus according to the present invention is not limited to a mobile phone terminal, and may be any information device provided with a function for identifying and authenticating an individual. In particular, a small dedicated authentication device that can be replaced with a conventional large personal authentication device and is authenticated only by a user's hand, or a portion responsible for authentication processing in the above embodiment (palmprint sensor 21, image recognition unit 22). The present invention can also be implemented in the form of a small authentication device that can be incorporated into other devices in which the image selection unit 23, the image conversion unit 24, the collation unit 25, and the palm print DB 26) are modularized.

In addition to the optical sensor described above, the palm print sensor 21 is composed of a capacitance sensor, that is, a large number of minute capacitors arranged in a matrix, and “wrinkles” or the like when the palm is brought into contact. A type of sensor that detects a change in capacitance caused by reflecting the unevenness may be used. Furthermore, an image of a palm print may be directly taken with a CCD (electrically coupled device) camera.
In the above embodiment, the palm print sensor 21 has been described as having a size of about two-thirds of that of the mobile phone terminal housing. However, this is one guideline and is approximately 3 to 5 cm square from the palm. It is desirable that the sensor size be capable of collecting the images.

It is a block diagram which shows the structure of the palm print authentication apparatus by one Embodiment of this invention. It is a figure which shows the positional relationship of the palm print sensor in the palm print authentication apparatus of FIG. 1, and the sensor and a palm. It is an example of the image of the palm collected by the palm print sensor of FIG. It is a flowchart which shows operation | movement of the palm print authentication apparatus of FIG. It is an example of the palm image extract | collected in the conventional palm print authentication method. It is an example of palm print data that is edge-extracted from a collected palm image.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Palmprint authentication apparatus 10 ... Telephone function part 21 ... Palmprint sensor 22 ... Image recognition part 23 ... Image selection part 24 ... Image conversion part 25 ... Collation part 26 ... Palmprint DB

Claims (5)

A sensor for collecting an image including a palm print from the palm of the hand,
Image recognition means for detecting, as a reference point, a valley portion between the middle finger and the base of the ring finger from the collected image;
In the image, image selection means for selecting a square image that is a line segment that passes through the reference point and that is parallel to the short side of the image and has a line segment having the same length as one side;
Image conversion means for extracting a palm print in the selected image by detecting an edge and converting the image into an image composed of edges;
Storage means for storing a matching image for a specific registered person;
A collation means for collating the converted image with the image for collation, and performing authentication when it is considered that both match based on a predetermined threshold;
A palmprint authentication apparatus comprising: On the computer,
Detecting a valley portion between the middle finger and the ring finger base from the input image including the palm print as a reference point;
In the image, selecting a square image that is a line segment passing through the reference point and having a line segment parallel to the short side of the image and having the same length as one side;
Extracting a palm print in the selected image by detecting an edge, and converting the image into an image composed of edges;
Collating the converted image with a collation image for a specific registered person stored in the storage unit;
As a result of the verification, authenticating when it is considered that the images match based on a predetermined threshold; and
A palmprint authentication program for executing Collecting an image including a palm print from a palm using a sensor;
Detecting by image recognition using the collected image as a reference point a valley portion between the middle finger and the base of the ring finger;
In the image, selecting a square image that is a line segment passing through the reference point and having a line segment parallel to the short side of the image and having the same length as one side;
Extracting a palm print in the selected image;
Collating the extracted palmprint with a palmprint of a specific registered person stored in the storage unit;
As a result of the collation, performing authentication when it is considered that the palm prints match based on a predetermined threshold;
A palmprint authentication method characterized by comprising: Detecting a valley portion between the middle finger and the base of the ring finger from an image including a palm print as a reference point by image recognition;
Extracting a square image having a line segment passing through the reference point and parallel to the short side of the image and having the same length as one side in the image;
A palmprint image extraction method characterized by comprising: A mobile phone terminal comprising the palm print authentication device according to claim 1,
The sensor in the palm print authentication apparatus is arranged on a surface opposite to a housing surface on which a main operation unit including a numeric keypad is provided.

Images