JP6167861B2 - Biological information extraction device, biological information extraction method, and biological information program - Google Patents

Description

  The present invention relates to a biological information extraction device, a biological information extraction method, and a biological information program.

  Patent Document 1 discloses a method of synthesizing one image from a vein image obtained by imaging the same living body part from a plurality of directions. Further, Patent Document 2 discloses a method of extracting a curved component and a directional component of a blood vessel pattern from a vein image and performing matching using these as feature amounts.

International Publication No. 2006/134669 JP 2007-249339 A

  When a three-dimensional linear pattern such as a palm vein or a finger vein is projected onto a two-dimensional image, the veins at the position of twist appear as complex areas including branches, intersections, etc. in the two-dimensional image. However, in the techniques of Patent Documents 1 and 2, when a position of a branch, an intersection, or the like in a two-dimensional image changes due to an influence of a posture change or the like when the user holds a living body part over a living body sensor, there is a place that should be connected originally It may be disconnected or a portion that is not originally connected may be connected, and it is not possible to appropriately extract the cutting and connection of the linear pattern. As a result, there has been a problem that authentication accuracy is lowered.

  In one aspect, the present invention provides a biological information extraction device that can improve the extraction accuracy of cutting, connecting, and the like of a linear pattern in a complex region including a branch, an intersection, and the like in a three-dimensional linear pattern of a living body Another object is to provide a biological information extraction method and a biological information program.

  In one aspect, the biological information extraction apparatus includes an acquisition unit that acquires a biological image obtained by imaging a three-dimensional linear pattern of a biological body, and an extraction unit that extracts the linear pattern from the biological image acquired by the acquisition unit. And a separation unit that separates the linear pattern extracted by the extraction unit into a plurality according to the characteristics of the linear pattern, and each linear pattern separated by the separation unit when the predetermined condition is satisfied A correction unit that connects the two to each other, and a synthesis unit that combines the linear patterns corrected by the correction unit.

  In a three-dimensional linear pattern in a living body, the extraction accuracy of the cutting and connection of the linear pattern can be improved even in a complicated region including branches, intersections, and the like.

(A) is a block diagram for demonstrating the hardware constitutions of the biometric information extraction apparatus which concerns on Example 1, (b) is a schematic diagram of a biosensor. (A) And (b) is an example of the biometric image obtained by the imaging of the biosensor. It is a block diagram of each function implement | achieved by execution of a biometric information extraction program. It is a flowchart showing an example of a registration process. (A) is an example of a part of vein pattern, (b)-(e) is an example of each component isolate | separated. It is a figure showing a correction process. (A) to (d) are vein patterns after correction processing, and (e) is a synthesized vein pattern. It is a flowchart showing an example of an authentication process. (A) is an example of a part of vein pattern, (b)-(e) is an example of each component isolate | separated. It is a flowchart showing an example of a registration process. (A) And (b) is a figure showing a correction process. It is a block diagram of each function implement | achieved by execution of a biometric information extraction program. It is a block diagram of each function implement | achieved by execution of a biometric information extraction program. The example of the vein pattern thinned by the thinning process is represented.

  Hereinafter, embodiments will be described with reference to the drawings.

  FIG. 1A is a block diagram for explaining a hardware configuration of the biological information extracting apparatus 100 according to the first embodiment. FIG. 1B is a schematic diagram of the biosensor 105 described later. Referring to FIG. 1A, the biological information extracting apparatus 100 includes a CPU 101, a RAM 102, a storage device 103, a display device 104, a biological sensor 105, and the like. Each of these devices is connected by a bus or the like.

  A CPU (Central Processing Unit) 101 is a central processing unit. The CPU 101 includes one or more cores. A RAM (Random Access Memory) 102 is a volatile memory that temporarily stores programs executed by the CPU 101, data processed by the CPU 101, and the like.

  The storage device 103 is a nonvolatile storage device. As the storage device 103, for example, a ROM (Read Only Memory), a solid state drive (SSD) such as a flash memory, a hard disk driven by a hard disk drive, or the like can be used. The biometric information extraction program according to the present embodiment is stored in the storage device 103. The display device 104 is a liquid crystal display, an electroluminescence panel, or the like, and displays a processing result of the biological information extraction device 100 and the like.

  The biological sensor 105 is a sensor that acquires a biological image by imaging a three-dimensional linear pattern of a user's biological body. The biological image acquired by the biological sensor 105 is a two-dimensional image. Examples of the three-dimensional linear pattern include a blood vessel pattern. Examples include palm veins, finger veins, retinal blood vessel patterns, and the like. In the present embodiment, the biometric sensor 105 is, for example, a sensor that images a palm vein, such as a CMOS (Complementary Metal Oxide Semiconductor) camera. For example, the biosensor 105 can image a vein by using near infrared rays.

  FIG. 2A and FIG. 2B are examples of biological images obtained by imaging with the biological sensor 105. As illustrated in FIGS. 2A and 2B, a vein pattern that is a three-dimensional linear pattern is formed into a two-dimensional image. In a two-dimensional image, veins at twisted positions are projected as a complex area including branches, intersections, and the like. Further, the position of these branches, intersections, and the like fluctuate due to the influence of posture fluctuation when the user holds the biological part over the biological sensor 105. In this case, there is a possibility that the portions that should originally be connected are interrupted or the portions that are not originally connected are connected, so that it is not possible to appropriately extract the division or connection of the linear pattern. Therefore, in the present embodiment, the accuracy of extracting whether the linear pattern is connected or disconnected in a complicated region including a branch, an intersection, etc. in the three-dimensional linear pattern of the living body is improved.

  The biometric information extraction program stored in the storage device 103 is expanded in the RAM 102 so as to be executable. The CPU 101 executes a biometric information extraction program developed in the RAM 102. Thereby, each process by the biometric information extraction device 100 is executed.

  FIG. 3 is a block diagram of each function realized by executing the biometric information extraction program. As illustrated in FIG. 3, by executing the biometric information extraction program, the biometric information extraction device 100 causes the feature extraction unit 10, the separation processing unit 20, the correction processing unit 30, the synthesis processing unit 40, the registration processing unit 50, and the authentication process. The unit 60 and the database 70 function.

  The feature extraction unit 10 extracts a palm vein pattern from the biological image acquired by the biological sensor 105. The separation processing unit 20 separates the vein pattern into a plurality according to the feature of the vein pattern. As the characteristics of the vein pattern, the direction of the vein, the thickness of the vein, and the like can be used. The correction processing unit 30 performs correction processing on each vein pattern separated by the separation processing unit 20. The synthesis processing unit 40 synthesizes each vein pattern after the correction processing by the correction processing unit 30 into one image. When executing the registration process, the registration processing unit 50 registers data obtained from the image synthesized by the synthesis processing unit 40 as registration data in the database 70 in association with the user ID. At the time of executing the authentication process, the authentication processing unit 60 uses the data obtained from the image synthesized by the synthesis processing unit 40 as collation data, and collates with registered data registered in the database 70. Details of each part will be described below.

  FIG. 4 is a flowchart illustrating an example of the registration process. As illustrated in FIG. 4, the biometric sensor 105 acquires a biometric image of the user's palm (step S1). Next, the feature extraction unit 10 extracts a vein pattern from the biological image acquired by the biological sensor 105 (step S2). This vein pattern is a projection of a three-dimensional vein pattern onto a two-dimensional image.

  Next, the separation processing unit 20 separates the vein pattern components extracted in step S2 into D (≧ 2) pieces for each direction (step S3). FIG. 5A is an example of a part of the vein pattern extracted in step S2. As illustrated in FIG. 5A, each vein is represented by a line or a line segment. For example, the separation processing unit 20 sets the horizontal right direction of the image to 0 °, and separates the image into a 0 ° direction component, a 45 ° direction component, a 90 ° direction component, and a 135 ° direction component. Each component may have a range. For example, a component facing 0 ° ± 22.5 ° (FIG. 5B), a component facing 45 ° ± 22.5 ° (FIG. 5C), and 90 ° ± 22.5 You may isolate | separate into the component (FIG.5 (d)) which faces (degree), and the component (FIG.5 (e)) which faces 135 degrees +/- 22.5 degrees. Here, the separation is synonymous with extracting a specific component from a biological image and generating a plurality of images of the same size.

  Next, the correction processing unit 30 performs correction processing on each vein pattern separated by the separation processing unit 20. Specifically, the correction processing unit 30 extracts an end point where the vein is interrupted in each separated vein pattern (step S4). Next, the correction processing unit 30 determines whether to connect the end points. For example, as illustrated in FIG. 6, the correction processing unit 30 obtains the distance d between all the extracted end points, and determines whether the distance d is equal to or less than a predetermined threshold Th <b> 1 (step S <b> 1). S5). Note that the example of FIG. 6 is the 0 ° direction component of FIG.

When it is determined as “Yes” in step S5, the correction processing unit 30 connects the end points (step S6). If it is determined “No” in step S5, step S6 is not executed. As an example of the value of the threshold Th1, a fixed value of about 3 pixels may be used, or a dynamic value such as an average value of the thickness of veins including the end points may be used. The connection between the end points may be interpolated using, for example, the following formula (1). In the following formula (1), I (x, y) represents a pixel value at coordinates (x, y), and γ _i is a constant that satisfies Σγ _i = 1.

  FIG. 7A shows a vein pattern after the correction process of FIG. FIG. 7B shows a vein pattern after the correction process of FIG. FIG. 7C shows a vein pattern after the correction process of FIG. FIG. 7D shows a vein pattern after the correction process of FIG. In these examples, since the distance between some of the end points in FIG. 5B is small, the two end points are connected in FIG. 7A.

  After the processing in steps S5 to S6 is completed for all end points in all separated vein patterns, the composition processing unit 40 composes one biological correction image from the separated vein patterns (step S7). For example, the composition processing unit 40 may take OR of each separated vein pattern, and take an average value of each pixel in a region where veins are common to a plurality of vein patterns. FIG. 7E shows a vein pattern after synthesis. Next, the registration processing unit 50 registers the data obtained from the image combined in step S7 as registration data in association with the user ID in the database 70 (step S8). With the above process, the registration process ends.

  According to the present embodiment, by separating the vein pattern components for each direction, the vein patterns can be separated for each component having a similar direction. Since the end points that are close to each other in the separated vein pattern are connected to each other, even when there are complex regions such as branches and intersections, the vein pattern can be stably disconnected and connected. That is, the extraction accuracy of the vein pattern can be improved.

  FIG. 8 is a flowchart illustrating an example of the authentication process. As illustrated in FIG. 8, processes similar to steps S <b> 1 to S <b> 7 in FIG. 4 are performed from step S <b> 11 to step S <b> 17. Next, the authentication processing unit 60 collates the collation data obtained from the image synthesized in step S17 with each registered data registered in the database 70 (step S18). Specifically, the authentication processing unit 60 calculates the similarity between the verification data and each registered data. Next, the authentication processing unit 60 determines whether there is registration data whose similarity is equal to or greater than a threshold value (step S19). If it is determined “Yes” in step S19, the authentication processing unit 60 determines that authentication is successful (step S20). If it is determined “No” in step S19, the authentication processing unit 60 determines that authentication is unsuccessful (step S21). In the present embodiment, since the extraction accuracy of the vein pattern is improved, the authentication accuracy is also improved.

(Modification)
In the above example, the vein pattern components are separated for each direction, but other features may be used. For example, the vein pattern components may be separated according to thickness. For example, the separation processing unit 20 converts the vein pattern (FIG. 9A) into a component having a thickness of 6 pixels or more (FIG. 9B) and a component having a thickness of 3 pixels to less than 6 pixels (FIG. 9). (C)) and a component having a thickness of less than 3 pixels (FIG. 9D) may be separated. In this case, the vein pattern can be separated for each component having a similar thickness. Since the end points that are close to each other in the separated vein pattern are connected to each other, even when there are complex regions such as branches and intersections, the vein pattern can be stably disconnected and connected. That is, the extraction accuracy of the vein pattern can be improved.

  In the second embodiment, the coupling reliability C is considered when determining whether or not the end points can be connected. The joint reliability C is a scale that represents whether or not the end points should be connected. For example, it can be said that the reliability is higher as the distance between the end points is shorter and the angle difference of the veins at the end points is smaller. Therefore, in this embodiment, the distance between the end points and the angular difference of the veins at the end points are used as the coupling reliability C. FIG. 10 is a flowchart illustrating an example of a registration process according to the present embodiment. As illustrated in FIG. 10, steps S31 to S35 are the same as steps S1 to S5 in FIG.

When it determines with "Yes" in step S35, the correction | amendment process part 30 calculates the coupling | bonding reliability C of the target endpoints (step S36). For example, as illustrated in FIGS. 11A and 11B, the correction processing unit 30 calculates the coupling reliability C from the distance d between the end points and the angular difference θ of the veins at the end points. . For example, the coupling reliability C can be expressed by the following formula (2). Α and β are positive constants and represent the weights of the distance d and the angle difference θ. The coupling reliability C becomes larger as the distance d and the angle difference θ are smaller.
C = α / d + β (1 + sin θ) (2)

  Next, the correction processing unit 30 determines whether or not the calculated coupling reliability C is greater than or equal to the threshold value Th2 (step S37). When it is determined as “Yes” in step S37, the correction processing unit 30 connects the end points (step S38). The connection between the end points may be interpolated using the above equation (1). In addition, when it determines with "No" by step S35, step S36-S38 is not performed. If it is determined “No” in step S37, step S38 is not executed.

  After the processing in steps S35 to S38 is completed for all end points in all separated vein patterns, the composition processing unit 40 composes one biological correction image from the separated vein patterns (step S39). Next, the registration processing unit 50 registers the data obtained from the biometric correction image synthesized in step S39 as registration data in association with the user ID in the database 70 (step S40). With the above process, the registration process ends. At the time of the authentication process, the authentication processing unit 60 uses the data obtained from the biometric correction image obtained in step S39 as collation data and performs collation by the same process as in steps S18 to S21 in FIG.

  According to this embodiment, the separated vein patterns are connected with the endpoints having high coupling reliability. Therefore, even if there are complicated areas such as branches and intersections, the vein patterns can be cut and connected. It can be extracted stably. That is, the extraction accuracy of the vein pattern can be improved. Note that other parameters may be taken into account for the coupling reliability. For example, the coupling reliability may be increased as the difference in the thickness of the veins at the end points is smaller.

  In the third embodiment, binarization processing is performed when a vein pattern is extracted. FIG. 12 is a block diagram of each function realized by executing the biometric information extraction program according to the present embodiment. As illustrated in FIG. 12, the difference from FIG. 3 is that a binarization processing unit 80 is further provided. In the vein pattern extracted by the feature extraction unit 10, the binarization processing unit 80 generates a biological binarized image in which a region where a vein is present is “1” and a region where no vein is present is “0”. The separation processing unit 20 separates the biological binarized image into a plurality of components for each component according to the characteristics of the vein pattern.

  In the fourth embodiment, binarization processing and thinning processing are performed at the time of vein pattern extraction. The thinning process refers to a process in which all the extracted vein pattern thicknesses are set to 1. FIG. 13 is a block diagram of each function realized by executing the biometric information extraction program according to the present embodiment. As illustrated in FIG. 13, the difference from FIG. 12 is that a thinning processing unit 90 is further provided. The thinning processing unit 90 corrects the thickness of each vein of the vein pattern binarized by the binarization processing unit 80 to 1 pixel. FIG. 14 shows an example of a vein pattern thinned by thinning processing.

  In each of the above examples, a palm vein is used as a three-dimensional linear pattern, but the present invention is not limited to this. As a three-dimensional linear pattern other than the palm vein, a palm vein, finger vein, retinal blood vessel pattern, or the like can be used. Each unit in each of the above examples is realized by executing a program, but hardware such as a dedicated circuit may be used.

  In each of the above examples, the biometric sensor 105 functions as an acquisition unit that acquires a biometric image obtained by imaging a three-dimensional linear pattern of a living organism. The feature extraction unit 10 functions as an extraction unit that extracts a linear pattern from the biological image acquired by the acquisition unit. The separation processing unit 20 functions as a separation unit that separates the linear pattern extracted by the extraction unit into a plurality of pieces according to the feature of the linear pattern. In each linear pattern separated by the separation unit, the correction processing unit 30 functions as a correction unit that connects the end points when a predetermined condition is satisfied. The synthesis processing unit 40 functions as a synthesis unit that synthesizes the linear patterns corrected by the correction unit.

  Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and changes can be made within the scope of the gist of the present invention described in the claims. It can be changed.

DESCRIPTION OF SYMBOLS 10 Feature extraction part 20 Separation processing part 30 Correction processing part 40 Composition processing part 50 Registration processing part 60 Authentication processing part 70 Database 100 Biometric information extraction apparatus 105 Biometric sensor

Claims (10)

An acquisition unit that acquires a biological image obtained by imaging a three-dimensional linear pattern of the biological body;
An extraction unit for extracting the linear pattern from the biological image acquired by the acquisition unit;
A separation unit that separates the linear pattern extracted by the extraction unit into a plurality according to the characteristics of the linear pattern;
In each linear pattern separated by the separation unit, a correction unit that connects end points when a predetermined condition is satisfied,
A biometric information extraction apparatus comprising: a synthesizing unit that synthesizes the linear patterns corrected by the correction unit.   The biological information extraction apparatus according to claim 1, wherein the separation unit separates the linear pattern in accordance with a direction of a component constituting the linear pattern.   The biological information extracting apparatus according to claim 1, wherein the separation unit separates the linear pattern according to a thickness of a component constituting the linear pattern.   The correction unit connects the two end points when the distance between the two end points is equal to or less than a threshold value in each linear pattern separated by the separation unit. The biological information extraction device according to any one of the above.   In the linear patterns separated by the separation unit, the correction unit is configured such that the distance between two end points is equal to or less than the threshold value and the angle difference between the lines at the two end points is equal to or less than the threshold value. 5. The biological information extracting apparatus according to claim 4, wherein the two end points are connected.   The biological information extracting apparatus according to claim 1, wherein the linear pattern is a vein pattern. A binarization processing unit that performs binarization processing on the linear pattern extracted by the extraction unit;
The biological information extraction according to any one of claims 1 to 6, wherein the separation unit separates the binarized linear pattern into a plurality according to the characteristics of the linear pattern. apparatus.   8. The thinning processing unit according to claim 7, further comprising a thinning processing unit that sets the thickness of the linear pattern to 1 pixel with respect to the binarized linear pattern generated by the binarization processing unit. Biological information extraction device. Obtaining a biological image obtained by imaging a three-dimensional linear pattern of the biological body;
Extracting the linear pattern from the biological image;
Separating the linear pattern into a plurality according to the characteristics of the linear pattern;
In each separated linear pattern, when a predetermined condition is satisfied, a correction process for connecting the end points is performed,
A biometric information extraction method comprising combining the linear patterns after the correction processing. Obtaining a biological image obtained by imaging a three-dimensional linear pattern of the biological body;
Extracting the linear pattern from the biological image;
Separating the linear pattern into a plurality according to the characteristics of the linear pattern;
In each separated linear pattern, when a predetermined condition is satisfied, a correction process for connecting the end points is performed,
A biometric information extraction program for causing a computer to execute a process of synthesizing the linear patterns after the correction process.

Images