JP5634933B2 - Biometric authentication system for detecting pseudo fingers - Google Patents

Description

The present invention relates to a technique for recognizing artificial fingers created artificially other than a living body when performing authentication using finger veins.

Recently, vein authentication apparatuses that perform personal authentication using finger vein pattern images have become widespread. The vein authentication apparatus images hemoglobin in blood as a dark shadow pattern by irradiating a finger with infrared light, and registers the shadow as a finger vein pattern.
At the time of authentication, an image of the user's finger is captured, and the finger vein pattern desired to be authenticated is compared with a finger vein pattern registered in advance to perform personal authentication.
As finger vein patterns to be registered in advance, a ballpoint pen other than a living body can be registered, and authentication is also possible. As a method for preventing authentication fraud with artificial fingers other than these living bodies, there are detection based on pulse measurement in Patent Document 1, detection by contact between a finger and a device in Patent Document 2, and the like.

JP 2005-046234 A JP 2005-071118 A

The vein authentication device is adopted by many financial institutions as personal authentication. However, when artificially created pseudo fingers are registered as vein patterns, replication is easy and damage due to impersonation of others occurs. It has been proved that vein patterns can be registered and authenticated by artificially created artificial fingers. Therefore, it is necessary to specify that the vein pattern to be registered is a living body.
Examples of a method for specifying a living body include the following Patent Document 1 and Patent Document 2. These devices require a device other than the photographing device, and need to touch the device with a finger.

An object of the present invention is to specify whether an image pattern other than a living body is included from a vein pattern of a captured finger without using a device other than an imaging device, specify an image pattern other than a living body, and determine a vein pattern other than a living body. It is to provide a biometric authentication system that can prevent registration and prevent false authentication such as impersonation.

In order to achieve the above object, a biometric authentication system for detecting a pseudo finger according to the present invention is a biometric authentication system for performing identity verification using biometric information related to a vein of a human finger,
An imaging unit that acquires image information including the biological information from an authentication target, and a polygonal line vector image and a binarization of the acquired image information by a polygonal line vectorization process and a binary bitmap process. Image conversion means for converting to a bitmap image, and determination for determining whether the image information is image information acquired from a living body or an image acquired from a pseudo finger based on the polygonal line vector image and a binarized bitmap image and means,
In the case where the determination means has a number of polygonal lines and branch points of a polygonal line vector image of 1 or more and the binarized bitmap image is the same color continuously for a certain number of rows or columns, it is an image acquired from a living body. comprising determining means and said Rukoto.

In the present invention, the binarized bitmap processing of the polygonal line vectorization processing is performed on the photographed vein image, and the vein pattern other than the living body is specified by combining each processing. This prevents registration of the pseudo finger vein pattern. Further, in Patent Document 1 and Patent Document 2, a device other than the imaging device is required. However, in the present invention, the determination is made only with the finger vein pattern, so that no extra device is required.

It is explanatory drawing which shows the structure of a vein authentication apparatus. It is explanatory drawing which shows the infrared light irradiation position of a vein authentication apparatus, and the position of a camera. It is a flowchart which shows the process which registers a finger vein pattern. It is a flowchart which shows the blood vessel pattern extraction of a finger vein pattern, a line graph process, and a binarization bitmap process. It is the figure which showed the blood-vessel pattern made into a polygonal line vector, the number of broken lines, and the number of intersections. It is the figure which showed the blood vessel pattern made into the binarization bitmap. It is the figure which showed the finger vein pattern which image | photographed the 2nd joint from the 1st joint. It is a flowchart for specifying the blood vessel pattern of a pseudo finger. It is a flowchart for determining whether a blood vessel pattern is a straight line. It is the figure which showed the linear blood vessel pattern by a broken line. It is the figure which showed the linear blood vessel pattern by a binarization bitmap. It is a flowchart for determining whether the blood vessel pattern is connected. It is the figure which showed that there is no connection of the blood vessel by a binarization bitmap. It is the figure which showed that the blood vessel by a broken line became cyclic | annular.

Hereinafter, the actual form of the biometric authentication system according to the present invention will be described.
FIG. 1 is a system configuration diagram showing the overall configuration of the biometric authentication system.
The personal computer (PC) 101 includes a vein authentication device 109 for registering finger vein patterns. Here, the finger vein pattern is information including image information showing a blood vessel pattern of a finger vein obtained by imaging from a surface such as a ventral side of a finger of a registrant's body.
The PC 101 includes an I / F 102 that transmits and receives signals and information to and from the vein authentication device 109, a CPU 103 that controls operations of each component and performs various calculations, and the CPU 103 performs calculations and the like. Memory 104 for temporarily storing information such as the result of the test, HDD 105 for registering and storing information such as finger information based on the registrant's body, and operations for inputting or selecting character information by a person operating the PC And a display 107 that is a means for displaying the processing results, information, and data of the program executed by the CPU 103 on the screen, and a speaker 108 that outputs work sounds during operation of the PC 101.
The CPU 103 has a function of temporarily storing the image information transmitted from the vein authentication device 109 in the memory 104, extracting a finger vein pattern from the image information, performing various processes, and storing it in the HDD 105. In addition, the infrared light source 111 has a function of performing processing for adjusting the amount of infrared light emitted.

  The vein authentication device 109 includes a finger placement unit 110 that places a finger for authenticating a registrant's finger at a predetermined position, an infrared light source 111 that emits infrared light to the finger, and the infrared light source 111 emits infrared light. A camera 112 that captures the veins of the finger projected by irradiation, a photosensor 113 for detecting that the finger is placed on the finger placement unit 110, and analog data of an image captured by the camera 112 are converted into digital data. A / D converter 114 that performs conversion processing into the image, an image input unit 115 that generates image information of digital information using digital data converted by the A / D converter 114, and transmission from the CPU 103 via the I / F 102 A light emission control unit 116 that controls the operation of the infrared light source 111 according to the emitted light control signal is provided.

FIG. 2 shows a state in which infrared rays emitted from the infrared light source 111 are radiated from the left and right sides of the finger 117, but there is also an apparatus that irradiates the finger abdomen.
The operation of the biometric authentication system in the present embodiment will be described with reference to the drawings.
First, a process for registering finger information including a registrant's finger vein pattern will be described with reference to the flowchart shown in FIG.
When a person who operates the PC 101 operates the keyboard 106 to start the vein pattern registration program stored in the HDD 105, the CPU 103 is initialized to the initialization of the entire hardware of the PC 101 and temporary editing necessary for the vein pattern registration program. A process of assigning a value is performed (step 301). When the CPU 103 completes the initialization process, the process proceeds to a process for shooting. A finger desired to be registered as a finger vein pattern is placed on the finger rest 110 of the vein authentication device 109, and “Now shooting vein” is displayed on the display 107 (step 302), and photographing is performed.

The CPU 103 performs a light emission amount adjustment process, appropriately adjusts the light emission amount of the infrared light source 101, and captures an image of a clear blood vessel pattern of the finger (step 303). The CPU 103 stores the image information in the memory 104, corrects the inclination of the image stored in the memory 104, deletes a portion where the blood vessel is not copied, extracts a blood vessel pattern of the vein, and extracts the extracted blood vessel pattern of the vein 5 is subjected to matrix processing by the polygonal line vectorization of FIG. 5 and the binary bit map of FIG.
Using the converted blood vessel pattern, a process for determining whether the blood vessel pattern is taken from a pseudo finger other than the living body is performed (step 305). If it is determined in the determination process that the finger is not a pseudo finger, the captured finger vein pattern is registered in the HDD 105 (step 307). A message “registration complete” or the like to end the processing is displayed on the display 107 (step 308), and the processing is ended.
If it is determined that the finger is a pseudo-finger, “Please retry registration” or the like is displayed on the display 107 as a message prompting the user to re-register (Step 309). If redo is selected, processing is performed from the beginning. If end is selected, nothing is registered in the HDD 105 and the vein pattern registration program is terminated.

Next, blood vessel pattern conversion will be described with reference to the flowchart of FIG.
FIG. 4 is a flowchart of processing for applying a polygonal line vector to a blood vessel pattern and processing for converting to a binary bit map and forming a matrix.
First, the tilt of the photographed image is corrected and corrected in the horizontal direction (step 401). Subsequently, a blood vessel portion is extracted from the corrected image (step 402). The extracted blood vessel is subjected to a process of forming a broken line using a line segment approximation for the blood vessel curve (step 403). A tracking start point is designated for a blood vessel that has been subjected to a broken line process, and an automatic tracking process is performed for the line segment, and the positions and number of broken lines and intersections are grasped. When performing automatic tracking of line segments, it is only necessary to extract the point where the intersection is divided into multiple line segments, but when determining the position of the broken line, check whether each line segment has a corner point To extract the number of vertices of the broken line and grasp the number (step 404). The processing result is temporarily stored in the memory 104.

FIG. 5 is a diagram showing the blood vessel pattern converted into a polygonal line vector, the number of broken lines, and the number of intersections.
Again, the extracted blood vessel is used to divide the blood vessel portion into black and the other blood vessels into white, so that a certain threshold t
Is converted into a white / black binary bitmap (step 405) to form a matrix (step 406). The processing result is temporarily stored in the memory 104.

FIG. 6 is a diagram showing a blood vessel pattern as a binary bitmap.
FIG. 7 is an example in which a vein from the first joint to the second joint of the finger is photographed. In the case of a blood vessel of a living body, the shape of the blood vessel is finely branched like a branch of a tree, and since it is a tube through which blood flows, there is a start point as an entrance and an end point as an exit. The photographed image is between the joints of the fingers, and in this description, the second joint side is the start point side.

FIG. 8 is a flowchart showing a process for omitting a pattern that is not suitable as a blood vessel in order to specify a blood vessel pattern of a pseudo finger.
The first condition for determining the pseudo-finger blood vessel pattern is that since a real blood vessel is usually finely branched, it is determined whether or not the blood vessel is a straight line having a length within a certain threshold range (step 801). . If it is not a straight line, it is determined whether or not two blood vessel patterns are running in parallel based on the possibility of using a ballpoint pen or the like (step 803).
If it is not a straight line pattern exceeding a certain threshold and the two blood vessel patterns do not run in parallel, it is determined that the corresponding blood vessel pattern is a blood vessel, and the process ends.
If it does not correspond, it is identified as a pseudo finger, and the pattern determination is terminated without performing the next determination. The second condition is that blood flows through the blood vessels, so the blood vessels are always connected, and the blood vessels cannot start in the middle. Therefore, it is determined that the blood vessel pattern is connected (step 803). If this condition is met, the finger is determined to be a finger, and the pattern determination process is terminated. If it does not correspond, it is identified as a pseudo finger and the pattern determination is terminated.

FIG. 9 is a flowchart for determining whether the blood vessel pattern is a straight line as the pseudo finger determination 1.
When the number of fold lines and the number of intersections of the bent blood vessel pattern is 1 or more (step 901), it can be determined that the line is branched or branched and is not a straight line.
Further, binarization processing is performed on the blood vessel pattern determined to have one or more broken lines and intersections.
It is determined whether or not the binarized image has continuous colors such as rows or columns (step 902). If it is a blood vessel pattern, it is determined that the same color is a blood vessel since the same color should have a continuous length (step 903).
If it is not continuous, it is assumed that the ink in the ballpoint pen has been formed in the middle, so it is determined that the blood vessel pattern is a pseudo finger (step 904).

FIG. 10 is a diagram showing a straight blood vessel pattern by broken lines.
It can be determined that the binarized bitmap matrix diagram is not a straight line when the row unit or the column unit is not continuously the same color (step 902).

FIG. 11 is a diagram illustrating a straight line in the case where the same color is continuously displayed in units of rows by the binarized bitmap.
When each condition is met, a blood vessel pattern imaged from a human finger is set (step 903). If not applicable, it is identified as a pseudo finger (step 904).

FIG. 12 is a flowchart for determining whether the blood vessel pattern is connected as the pseudo finger determination 2.
If there is black in any of the row directions of the matrix diagram (step 1201), it can be determined that a blood vessel exists.
When automatic tracking is performed from the start point of the bent blood vessel pattern to reach the intersection or end point (step 1202), it can be determined that blood is flowing (step 1203).
If it does not reach the intersection or end point and passes through the start point again, it can be determined that the line is circular and not a blood vessel (step 1204).

  FIG. 13 and FIG. 14 are diagrams showing that the third row in the binarized bitmap is all white and no blood vessels are connected.

FIG. 14 is a diagram showing that the blood vessel formed by the broken line is annular.
The figure shown here is an example of an image that is read as to whether or not the ballpoint pen can be recognized as a circular slice.
In this case, a structure such as a circle or a rectangle in which the start point and the end point are continuous is not applicable to the blood vessel pattern.
As described above, in the present embodiment, in a biometric authentication system that performs personal identification using biometric information related to a human finger vein, image information including biometric information is acquired from an authentication target, and the acquired image information is The image information is converted into a polygonal line vector image and a binary bitmap image by performing a polygonal line vectorization process and a binary bitmap process, and the image information is based on the polygonal line vector image and the binary bitmap image. Is an image information acquired from a living body or an image acquired from a pseudo finger.
For this reason, it becomes possible to detect a pseudo finger or the like without adding a special device.

101 PC
102 I / F
103 CPU
104 Memory 105 HDD
106 Keyboard 107 Display 108 Speaker 109 Vein authentication device 110 Finger rest 111, 202 Infrared light source 112, 203 Camera 113 Photo sensor 114 A / D converter 115 Image input unit 116 Light emission control unit 117, 201, 701 Finger 702 Finger vein pattern Pattern extracted from blood vessels

Claims (1)

A biometric authentication system that performs identity verification using biometric information related to a human finger vein,
An imaging unit that acquires image information including the biological information from an authentication target, and a polygonal line vector image and a binarization of the acquired image information by a polygonal line vectorization process and a binary bitmap process. Image conversion means for converting to a bitmap image, and determination for determining whether the image information is image information acquired from a living body or an image acquired from a pseudo finger based on the polygonal line vector image and a binarized bitmap image and means,
In the case where the determination means has a number of polygonal lines and branch points of a polygonal line vector image of 1 or more and the binarized bitmap image is the same color continuously for a certain number of rows or columns, it is an image acquired from a living body biometric authentication system for detecting pseudo finger, characterized in Rukoto comprises determining means.

Images