JP4760422B2 - Biometric identification device - Google Patents

Description

  The present invention relates to a biometric identification device that identifies a living body by photographing a blood vessel of the living body.

  As a biometric identification device for imaging a living blood vessel and identifying the living body, at the time of registration, the blood vessel of the living body is photographed two-dimensionally to obtain a two-dimensional blood vessel image, and the obtained two-dimensional blood vessel image is analyzed and the blood vessel is analyzed. Two-dimensional feature points of the pattern are registered as feature points for two-dimensional matching, and at the time of matching, a two-dimensional blood vessel image is obtained by photographing a blood vessel of a living body two-dimensionally, and the obtained two-dimensional blood vessel image is analyzed. Then, the two-dimensional feature points of the blood vessel pattern are extracted, and the biological object is identified by comparing the two-dimensional feature points of the previously registered blood vessel pattern with the two-dimensional feature points of the extracted blood vessel pattern. What to do is offered.

  By the way, in the registration and verification using the two-dimensional feature points of the blood vessel pattern in this way, if the posture of the living body is different between the registration time and the verification time, the arrangement of the two-dimensional matching feature points of the blood vessel pattern at the time of registration is performed. Since the mode and the arrangement mode of the two-dimensional feature points of the blood vessel pattern at the time of collation are different, it is difficult to properly collate both, and there is a problem that the accuracy of collation is lowered.

Therefore, at the time of registration, a three-dimensional blood vessel image is acquired by three-dimensionally imaging a blood vessel of a living body to obtain a three-dimensional blood vessel image or by synthesizing a plurality of two-dimensional blood vessel images acquired by two-dimensional imaging. The acquired 3D blood vessel image is analyzed and the 3D feature points of the blood vessel pattern are registered as 3D matching feature points, and the blood vessels of the living body are photographed 3D at the time of matching. A blood vessel image is acquired, or a plurality of two-dimensional blood vessel images acquired by two-dimensional imaging are combined to acquire a three-dimensional blood vessel image, and the acquired three-dimensional blood vessel image is analyzed to analyze 3 of the blood vessel pattern. There is also provided one that extracts a three-dimensional feature point and collates a pre-registered three-dimensional matching feature point of a blood vessel pattern with the three-dimensional feature point of the extracted blood vessel pattern (see, for example, Patent Document 1). ).
JP 2001-273497 A

  However, in the configuration described in Patent Document 1 described above, a biological blood vessel is photographed three-dimensionally or two-dimensionally in order to extract a three-dimensional feature point of a blood vessel pattern not only at the time of registration but also at the time of collation. Since it is necessary to synthesize a plurality of two-dimensional blood vessel images acquired by photographing, there are problems that the matching process becomes complicated and the apparatus becomes large. In particular, in a configuration in which a blood vessel of a living body is imaged three-dimensionally, it is difficult to image the living body in a free space, so there is a problem that it is inconvenient for both a collator to be collated and a collated person to be collated. is there.

  The present invention has been made in view of the above-described circumstances, and the purpose thereof is that collation accuracy does not decrease, collation processing is not complicated, and the apparatus is increased in size. In addition, it is an object of the present invention to provide a biometric identification device that can appropriately identify a living body while improving convenience for both a collator to be collated and a collated person to be collated.

According to the first aspect of the present invention, when a blood vessel of a living body is photographed two-dimensionally and a two-dimensional blood vessel image is acquired by the photographing means, the two-dimensional feature point extracting means analyzes the two-dimensional blood vessel image. Then, a two-dimensional feature point of the blood vessel pattern is extracted. In addition, the change detecting means sets the posture of the living body at the time when the blood vessel image that is the origin of the three-dimensional feature point of the blood vessel pattern is acquired as the basic posture, and at the time when the two-dimensional blood vessel image is acquired by the imaging means. The posture of the two-dimensional blood vessel image corresponding to the distance from the predetermined position of the biometric identification device to the reference plane and the inclination of the reference plane detected by the imaging means is set as the matching posture, and the degree of change from the basic posture to the matching posture Is detected.

The two-dimensional matching feature point generation means uses at least three points belonging to the photographing target at the time of photographing as reference points, measures the distance from the predetermined position of the biometric identification device to each reference point, and uses the reference of the photographing target at the time of photographing. When the distance from the predetermined position of the biometric identification device to the reference plane and the inclination of the reference plane are detected by the imaging unit, the distance from the predetermined position of the biometric identification device detected by the imaging unit to the reference plane A cut-out plane corresponding to the inclination of the reference surface and the degree of change from the basic posture detected by the change detecting means to the posture at the time of matching is generated. Further, the two-dimensional matching feature point generation means acquires a three-dimensional feature point of the blood vessel pattern when the living body is in the basic posture from the three-dimensional feature point storage means, and detects a change among the three-dimensional feature points of the blood vessel pattern. A three-dimensional feature point corresponding to the degree of change detected by the means is cut out and plotted on a plane to generate a two-dimensional matching feature point of the blood vessel pattern. Then, the collating unit collates the two-dimensional matching feature point of the blood vessel pattern generated in this way with the two-dimensional feature point of the blood vessel pattern extracted in this way.

  Thereby, by collating using the two-dimensional feature point of the blood vessel pattern, it is not necessary to photograph the blood vessel of the living body three-dimensionally at the time of collation, and it is sufficient to photograph the blood vessel of the living body two-dimensionally. It is possible to take a picture in free space, to improve convenience for both the collator to be collated and the collated person to be collated, to avoid the complexity of the collation process, and to the apparatus It is possible to avoid an increase in size. Furthermore, by selecting a three-dimensional feature point of a blood vessel pattern and generating a two-dimensional matching feature point, it is possible to avoid a reduction in matching accuracy and to appropriately identify a living body.

  According to the second aspect of the present invention, the distortion detecting means detects the degree of distortion from the basic posture to the matching posture, and the two-dimensional matching feature point generating means is a blood vessel pattern corresponding to the degree of distortion. 2D matching feature points are generated. Thereby, even when the posture at the time of matching of the living body is distorted from the basic posture, the two-dimensional matching feature point of the blood vessel pattern corresponding to the degree of the distortion is matched with the two-dimensional feature point of the blood vessel pattern. As a result, the distortion from the basic posture to the posture at the time of collation can be corrected and collation can be performed, and the accuracy of collation can be improved.

  According to the invention described in claim 3, the distortion detecting means detects the degree of distortion from the basic posture to the matching posture, and the two-dimensional blood vessel image correcting means corresponds to the degree of distortion, and the two-dimensional blood vessel The image is corrected, and the two-dimensional feature point extracting means analyzes the corrected two-dimensional blood vessel image and extracts a two-dimensional feature point of the blood vessel pattern. As a result, even when the posture at the time of matching of the living body is distorted from the basic posture, the blood vessel obtained by analyzing the two-dimensional matching feature point of the blood vessel pattern and the two-dimensional blood vessel image corrected in accordance with the degree of the distortion. By collating with the two-dimensional feature points of the pattern, it is possible to perform the collation by correcting the distortion from the basic posture to the posture at the time of collation in the same manner as described in claim 2 above. Can be increased.

  According to the invention described in claim 4, when the outline of the living body is detected by the outline detecting means, the enlarging / reducing means converts the 2D blood vessel image based on the detection result to the origin of the 3D feature point of the blood vessel pattern. The two-dimensional feature point extracting means extracts the two-dimensional feature point of the blood vessel pattern by analyzing the enlarged / reduced two-dimensional blood vessel image. Thereby, the two-dimensional matching feature point of the blood vessel pattern and the two-dimensional feature point of the blood vessel pattern can be collated at substantially the same magnification, and the collation accuracy can be improved.

  According to the fifth aspect of the present invention, the two-dimensional matching feature point generating means generates three-dimensionally each of the two-dimensional matching feature points of the blood vessel pattern. As a result, redundancy can be imparted to the two-dimensional matching feature points of the blood vessel pattern, and matching accuracy can be improved.

  Hereinafter, as an embodiment of the present invention, a case where a living body is identified by collating a blood vessel (vein) of a palm or a finger will be described with reference to the drawings. FIG. 1 is a functional block diagram of a registration scanner used when registering feature points of a blood vessel pattern of a palm or a finger and a handy scanner for verification used when checking feature points of a blood vessel pattern of a palm or a finger. Is shown.

  The registration scanner 1 is a fixed scanner that is fixedly installed on, for example, a wall or a desk, and includes a control unit 2, an imaging unit 3, an image processing unit 4, a 3D image generation unit 5, and a 3D feature point generation unit 6. And a communication unit 7. The control unit 2 includes a CPU, a RAM, a ROM, and the like, and controls the overall operation of the registration scanner 1 by controlling the operations of the function block units 3 to 7.

  The imaging unit 3 includes a light source 8 and a sensor 9. When a shooting command is input from the control unit 2, the light source 8 emits light, and the light that is emitted from the light source 8 and reflected by the blood vessels of the palm and fingers is reflected. The sensor 9 receives the palm and finger blood vessels two-dimensionally to obtain a two-dimensional blood vessel image. The photographing unit 3 photographs the palm and finger blood vessels in a predetermined basic posture by photographing the blood vessels of the palm and fingers while the posture of the palm and fingers is fixed at a predetermined position.

  When the 2D blood vessel image acquired by the imaging unit 3 is input from the imaging unit 3 (sensor 9), the image processing unit 4 performs image processing on the input 2D blood vessel image according to an image processing algorithm to generate a 3D image. Output to unit 5. When a plurality of two-dimensional blood vessel images subjected to image processing are input from the image processing unit 4, the three-dimensional image generation unit 5 generates a three-dimensional blood vessel image by synthesizing the input two-dimensional blood vessel images. Output to the feature point generator 6. When the 3D feature point generation unit 6 receives a 3D blood vessel image from the 3D image generation unit 5, the 3D feature point generation unit 6 analyzes the input 3D blood vessel image and generates a 3D feature point of the blood vessel pattern.

  When the communication unit 7 receives a storage command from the control unit 2, the communication unit 7 stores the three-dimensional feature points of the blood vessel pattern generated by the three-dimensional feature point generation unit 6 in a three-dimensional feature point database 31 (three-dimensional feature point storage means in the present invention). ) To be stored (stored).

  The collation handy scanner 11 is structurally different from the registration scanner 1 described above, and is a portable scanner that can be easily carried by the user, and includes a control unit 12 and a photographing unit 13 (in the present invention). Imaging means, outer shape detection means), image processing section 14 (two-dimensional blood vessel image correction means, enlargement / reduction means in the present invention), two-dimensional feature point extraction section 15 (two-dimensional feature point extraction means in the present invention), A change detection unit 16 (change detection unit and distortion detection unit in the present invention), a communication unit 17, a two-dimensional matching feature point generation unit 18 (two-dimensional matching feature point generation unit in the present invention), and a matching unit 19 ( It is provided with collating means in the present invention.

  The control unit 12 includes a CPU, a RAM, a ROM, and the like, and controls the overall operation of the collation handy scanner 11 by controlling the operations of the function block units 13 to 19. The imaging unit 13 includes a light source 20 and a sensor 21. When an imaging command is input from the control unit 12, the light source 20 emits light, and the light that is emitted from the light source 20 and reflected by a blood vessel of a palm or a finger is reflected. By receiving the sensor 21, the blood vessel of the palm or finger is photographed two-dimensionally to obtain a two-dimensional blood vessel image.

  Since the collation handy scanner 11 is a portable scanner, the photographing unit 13 creates a reference plane for a palm or a finger at the time of photographing, so that a predetermined position (for example, the light source 20) of the collation handy scanner 11 is created. ) To the reference plane and the inclination of the reference plane, as well as the palm and finger contours at the time of shooting. In this case, when collating the palm, as shown in FIG. 2 (a), the photographing unit 13 uses the base of the thumb, the base of the index finger and the middle finger, and the base of the ring finger and the little finger as reference points. A reference plane is created by measuring the distance from a predetermined position of the handy scanner 11 for verification to each reference point. When collating the finger, the photographing unit 13 uses the three points of the base of the index finger and the middle finger, the tip of the index finger, and the tip of the middle finger as reference points, as shown in FIG. A reference plane is created by measuring the distances from 11 predetermined positions to each reference point.

  When a two-dimensional blood vessel image is input from the imaging unit 13 (sensor 21), the image processing unit 14 performs image processing on the input two-dimensional blood vessel image according to an image processing algorithm, and at the time of registering the input two-dimensional blood vessel image. The image is enlarged / reduced to substantially match the size of the two-dimensional blood vessel image and output to the two-dimensional feature point extraction unit 15 and the change detection unit 16. The two-dimensional feature point extraction unit 15 receives the image processing and the enlarged / reduced two-dimensional blood vessel image from the image processing unit 14, analyzes the input two-dimensional blood vessel image, and extracts the two-dimensional feature points of the blood vessel pattern. And output to the collation unit 19.

  The change detection unit 16 includes a two-dimensional blood vessel image acquired by the imaging unit 13 from the imaging unit 13 (sensor 21), a distance from the predetermined position of the collation handy scanner 11 detected by the imaging unit 13, and a reference plane. Is input, the orientation of the two-dimensional blood vessel image corresponding to the input distance from the predetermined position of the verification handy scanner 11 to the reference plane and the inclination of the reference plane is set as the verification attitude, and the basic attitude at the registration described above. The degree of change to the matching posture is detected, and the detection result is output to the two-dimensional matching feature point generation unit 18.

  When receiving a reading command from the control unit 12, the communication unit 17 inputs and reads the three-dimensional feature points of the blood vessel pattern stored in the three-dimensional feature point database 31 from the three-dimensional feature point database 31, and the read blood vessel The three-dimensional feature points of the pattern are output to the two-dimensional matching feature point generation unit 18. When the two-dimensional matching feature point generation unit 18 inputs the three-dimensional feature point of the blood vessel pattern from the communication unit 17 and the detection result from the change detection unit 16, the three-dimensional feature point of the input blood vessel pattern Are selected from the three-dimensional feature points corresponding to the detection result of the change detection unit 16 to generate a two-dimensional matching feature point and output it to the matching unit 19.

  When the matching unit 19 inputs the two-dimensional matching feature point of the blood vessel pattern from the two-dimensional matching feature point generation unit 18 and inputs the two-dimensional feature point of the blood vessel pattern from the two-dimensional feature point extraction unit 15, both of them That is, based on the two-dimensional matching feature point of the blood vessel pattern based on the two-dimensional blood vessel image of the palm or finger photographed by the registration scanner 1 and the two-dimensional blood vessel image of the palm or finger photographed by the handy scanner 11 for matching. The two-dimensional feature point of the blood vessel pattern is collated.

  In the above configuration, the three-dimensional feature point database 31 is composed of, for example, an IC card, and the registration scanner 1 is in a state where the IC card is normally mounted in a mounting slot (not shown). The 3D feature point of the blood vessel pattern based on the 2D blood vessel image of the palm or finger can be output to the IC card and stored. The handy scanner 11 for collation has an IC card installed in a slot (not shown). 3), the three-dimensional feature points of the blood vessel pattern stored in the IC card can be input and read from the IC card. Note that the 3D feature point database 31 is not limited to the IC card as described above, and may be configured of, for example, a server. The registration scanner 1 transmits the 3D feature points of the blood vessel pattern through the network. The stored handy scanner 11 may store the three-dimensional feature points of the blood vessel pattern stored in the server through a network.

  Next, as an operation of the above-described configuration, the collation handy scanner 11 executes in parallel the processing for generating the two-dimensional collation feature points of the blood vessel pattern and the processing for extracting the two-dimensional feature points of the blood vessel pattern, and thereafter The case of executing the process of collating both will be described with reference to FIGS.

  In the processing for generating the two-dimensional matching feature point of the blood vessel pattern, the matching handy scanner 11 has three points of the distance from the predetermined position of the matching handy scanner 11 to the reference plane and the inclination of the reference plane in the imaging unit 13. Is detected based on the reference point (step S1), and the change detection unit 16 detects the degree of change from the basic posture to the matching posture (step S2).

  Next, as shown in FIG. 4A, the collation handy scanner 11 uses the two-dimensional collation feature point generation unit 18 to detect the distance from the predetermined position of the collation handy scanner 11 to the reference plane and the reference plane. A cut-out plane corresponding to the inclination of the surface and the degree of change from the basic posture to the posture at the time of collation is generated (step S3), and the distance from the predetermined position of the handy scanner 11 for collation to the reference surface and the light source 20 Based on the light intensity, the thickness of the cut plane is calculated (step S4).

  Next, the matching handy scanner 11 inputs and reads the three-dimensional feature points (see FIG. 4B) of the blood vessel pattern from the three-dimensional feature point database 31 through the communication unit 17 (step S5). Then, the collation handy scanner 11 cuts out the three-dimensional feature points of the inputted blood vessel pattern in the two-dimensional collation feature point generation unit 18 as shown in FIG. 2D matching feature points are generated (step S6). At this time, the matching handy scanner 11 generates each of the two-dimensional matching feature points of the blood vessel pattern three-dimensionally (as a spherical point having a volume).

  On the other hand, in the processing for extracting the two-dimensional feature point of the blood vessel pattern, the collation handy scanner 11 acquires a two-dimensional blood vessel image by the photographing unit 13 (step S7), and performs a reference from a predetermined position of the collation handy scanner 11. The distance to the surface is detected based on the three reference points (step S8), and the image processing unit 14 makes the acquired two-dimensional blood vessel image substantially coincide with the size of the two-dimensional blood vessel image at the time of registration. Enlarge / reduce (step S9). Then, the matching handy scanner 11 analyzes the two-dimensional blood vessel image at the two-dimensional feature point extraction unit 15 and extracts the two-dimensional feature points of the blood vessel pattern (step S10).

  Then, the collation handy scanner 11 collates the two-dimensional feature point of the blood vessel pattern generated in this way with the two-dimensional feature point of the blood vessel pattern extracted in this manner in the collation unit 19 ( In step S11), when it is detected that both arrangement modes match (“YES” in step S12), the collation result is determined to be positive (step S13), while both arrangement modes match. If it is detected that it is not (“NO” in step S12), it is determined that the collation result is negative (step S14), and the series of processing ends. Note that the criterion for determining whether or not the arrangement modes of the two are the same is determined according to the required accuracy.

  Through the series of processes described above, the matching handy scanner 11 selects a three-dimensional feature point corresponding to the degree of change from the basic posture to the matching posture from the three-dimensional feature points of the blood vessel pattern generated at the time of registration. A feature point for two-dimensional matching is generated, a two-dimensional blood vessel image acquired at the time of matching is analyzed to extract a two-dimensional feature point of the blood vessel pattern, and a two-dimensional matching feature point and a two-dimensional feature point of the blood vessel pattern are extracted. And the correctness of the verification result is confirmed.

  By the way, the above describes the case where it is assumed that the posture at the time of collation of the palm or finger is not distorted from the basic posture, but it is assumed that the posture at the time of collation of the palm or finger is distorted from the basic posture after this. The case will be described.

  In this case, the imaging unit 13 creates a reference plane by increasing the reference points. Specifically, in the case of collating the palm, the photographing unit 13, as shown in FIG. 6A, the base of the thumb, the base of the index finger and the middle finger, the base of the ring finger and the little finger, the base of the thumb and the ring finger 5 points of reference point, which is point-symmetrical to the base of the index finger and the middle finger centered on the intermediate position between the base of the thumb and the little finger, and the intermediate position of the base of the thumb, the ring finger and the little finger The distance from the predetermined position of the handy scanner 11 to each reference point is measured to create a reference plane. Further, when collating fingers, as shown in FIG. 6 (b), the photographing unit 13 includes a base between the index finger and the middle finger, a tip of the index finger, a tip of the middle finger, a base of the index finger and the middle finger, and a tip of the index finger. The reference plane is determined by measuring the distance from the predetermined position of the handy scanner 11 for collation to each reference point, with the five points of the position between the base of the index finger and the middle finger and the tip of the middle finger as reference points. Create

  The change detection unit 16 includes a two-dimensional blood vessel image acquired by the imaging unit 13 from the imaging unit 13 (sensor 21), a distance from the predetermined position of the collation handy scanner 11 detected by the imaging unit 13, and a reference plane. Is input, the orientation of the two-dimensional blood vessel image corresponding to the input distance from the predetermined position of the verification handy scanner 11 to the reference plane and the inclination of the reference plane is set as the verification attitude, and the basic attitude at the registration described above. The degree of change from the position to the matching posture is detected and the degree of distortion is also detected.

  Here, as a method of correcting the degree of distortion from the basic posture to the posture at the time of matching, a method of correcting by a process of generating a two-dimensional matching feature point of the blood vessel pattern and a two-dimensional feature point of the blood vessel pattern are extracted. There are two methods, a method of correcting by processing.

First, a method for correcting the degree of distortion from the basic posture to the posture at the time of matching in the process of generating the two-dimensional matching feature points of the blood vessel pattern will be described with reference to FIG.
In the processing for generating the two-dimensional matching feature point of the blood vessel pattern, the matching handy scanner 11 has five points of the distance from the predetermined position of the matching handy scanner 11 to the reference plane and the inclination of the reference plane in the imaging unit 13. Is detected based on the reference point (step S21), and the change detection unit 16 detects the degree of change from the basic posture to the collation posture (step S22), and here, from the basic posture to the collation posture. The degree of distortion is also detected (step S23).

  Next, the matching handy scanner 11 uses the two-dimensional matching feature point generation unit 18 to perform the matching posture from the detected distance from the predetermined position of the matching handy scanner 11 to the reference surface, the inclination of the reference surface, and the basic posture. A cut curved surface corresponding to the degree of change and the degree of distortion is generated (step S24), and based on the distance from the predetermined position of the hand-held scanner 11 for collation to the reference plane and the intensity of light cast from the light source 20. Then, the thickness of the cut curved surface is calculated (step S25).

  Next, the matching handy scanner 11 inputs and reads the three-dimensional feature points of the blood vessel pattern from the three-dimensional feature point database 31 in the communication unit 17 (step S26). Then, the matching handy scanner 11 generates a two-dimensional matching feature point of the blood vessel pattern by using the two-dimensional matching feature point generation unit 18 to cut out and plot the three-dimensional feature point of the inputted blood vessel pattern on the curved surface. (Step S27).

  On the other hand, in the process of extracting the two-dimensional feature point of the blood vessel pattern, the matching handy scanner 11 detects the distance from the predetermined position of the matching handy scanner 11 to the reference plane based on the five reference points (step S28). ), Extracting a two-dimensional feature point of the blood vessel pattern. The collation handy scanner 11 uses the collation unit 19 to generate the two-dimensional collation feature points of the blood vessel pattern thus generated by correcting the degree of distortion from the basic posture to the collation posture. The two-dimensional feature points of the extracted blood vessel pattern are collated (steps S11 and S12), and whether the collation result is correct is determined (steps S13 and S14).

  Through the above-described series of processing, the matching handy scanner 11 corrects the degree of distortion from the basic posture to the posture at the time of matching in the processing for generating the two-dimensional matching feature point of the blood vessel pattern, and the degree of distortion is corrected. The two-dimensional matching feature point of the blood vessel pattern is matched with the two-dimensional feature point of the blood vessel pattern, and the correctness of the matching result is determined.

Next, a method for correcting the degree of distortion from the basic posture to the collation posture in the process of extracting the two-dimensional feature points of the blood vessel pattern will be described with reference to FIG.
In this case, in the process of extracting the two-dimensional feature point of the blood vessel pattern, the collation handy scanner 11 acquires a two-dimensional blood vessel image by the imaging unit 13 (step S31), and starts from a predetermined position of the collation handy scanner 11. The distance to the reference plane is detected based on the five reference points (step S32). Here, the change detection unit 16 detects the degree of distortion from the basic posture to the matching posture (step S33). Next, the matching handy scanner 11 corrects the acquired two-dimensional blood vessel image in accordance with the degree of distortion in the image processing unit 14 (step S34), and the corrected two-dimensional blood vessel image is registered at the time of registration. Enlargement / reduction is performed so as to substantially match the size of the dimensional blood vessel image (step S35). Then, the matching handy scanner 11 analyzes the two-dimensional blood vessel image at the two-dimensional feature point extraction unit 15 and extracts the two-dimensional feature points of the blood vessel pattern (step S36).

  On the other hand, in the process of generating the two-dimensional matching feature point of the blood vessel pattern, the matching handy scanner 11 detects the distance from the predetermined position of the matching handy scanner 11 to the reference plane based on the five reference points ( Step S37), generating a two-dimensional matching feature point of the blood vessel pattern. Then, the collation handy scanner 11 corrects the two-dimensional collation feature points of the blood vessel pattern generated in this way and the degree of distortion from the basic posture to the collation posture in this way by the collation unit 19. The two-dimensional feature points of the extracted blood vessel pattern are collated (steps S11 and S12), and whether the collation result is correct is determined (steps S13 and S14).

  Through the series of processes described above, the matching handy scanner 11 corrects the degree of distortion from the basic posture to the matching posture by the process of extracting the two-dimensional feature points of the blood vessel pattern, and the two-dimensional matching feature points of the blood vessel pattern. And the two-dimensional feature point of the blood vessel pattern whose degree of distortion is corrected, and the correctness of the verification result is determined.

  As described above, according to the present embodiment, the matching handy scanner 11 is configured to match the two-dimensional matching feature point of the blood vessel pattern with the two-dimensional feature point of the blood vessel pattern. There is no need to photograph the blood vessels of the palm and fingers in three dimensions, and it is only necessary to photograph the blood vessels of the palm and fingers in two dimensions. The palm and fingers can be photographed in free space. It is possible to improve convenience for both of the persons to be verified, to avoid the complexity of the verification process, and to prevent the apparatus from becoming large. Furthermore, since the 3D feature point corresponding to the degree of change from the basic posture to the posture at the time of matching is selected from the 3D feature points of the blood vessel pattern, the feature point for 2D matching is generated. It is possible to avoid a decrease in accuracy, and it is possible to appropriately identify palms and fingers.

  Further, even when the collation posture of the palm or finger is distorted from the basic posture, the degree of distortion from the basic posture to the collation posture is detected, and the degree of the distortion is corrected to correct the two-dimensional blood vessel pattern. Since the verification feature point and the two-dimensional feature point of the blood vessel pattern are verified, the accuracy of the verification can be improved. In addition, the outer shape of the palm or finger is detected, and based on the detection result, the two-dimensional blood vessel image at the time of matching is enlarged or reduced so as to substantially match the size of the two-dimensional blood vessel image at the time of registration. Since the two-dimensional blood vessel image is analyzed to extract the two-dimensional feature points of the blood vessel pattern, the two-dimensional matching feature points of the blood vessel pattern and the two-dimensional feature points of the blood vessel pattern are collated at substantially the same magnification. It is possible to improve the accuracy of verification. Further, since each of the two-dimensional matching feature points of the blood vessel pattern is generated three-dimensionally, the two-dimensional matching feature points of the blood vessel pattern can be made redundant, and the accuracy of matching is improved. be able to.

The present invention is not limited to the above-described embodiment, and can be modified or expanded as follows.
The configuration may be applied to authentication for blood vessel patterns of parts other than the palm and fingers (for example, the back of the hand).
The registration scanner is configured to three-dimensionally capture a blood vessel of a palm or a finger to acquire a three-dimensional blood vessel image, and analyze the acquired three-dimensional blood vessel image to generate a three-dimensional feature point of the blood vessel pattern. You may do it. Moreover, the structure provided with the alerting | reporting means (For example, a display, LED, a speaker, etc.) which alert | reports the correctness of a collation result may be sufficient.

If it is assumed that the posture when collating palms and fingers is not distorted from the basic posture, a configuration may be adopted in which a reference plane is created based on four or more reference points. When it is assumed that the posture is distorted, a configuration in which a reference plane is created based on six or more reference points may be used.
When it is assumed that the posture at the time of collation of the palm or finger is distorted from the basic posture, the degree of distortion from the basic posture to the posture at the time of collation is corrected by the process of generating the feature points for two-dimensional matching of the blood vessel pattern A configuration in which correction is performed by a process of extracting a two-dimensional feature point of a blood vessel pattern may be used.

Functional block diagram showing an embodiment of the present invention Diagram showing the reference point when creating a reference plane for palms and fingers flowchart A diagram representing a cut-out plane (a), a diagram representing a three-dimensional feature point of a blood vessel pattern (b) The figure (a) showing the cut-out plane which plotted the three-dimensional feature point of the blood vessel pattern, The figure (b) showing the two-dimensional feature point of the blood vessel pattern 6 equivalent diagram 3 equivalent figure 3 equivalent figure

Explanation of symbols

  In the drawing, 11 is a handy scanner for comparison (biological identification device), 13 is an imaging unit (imaging unit, outer shape detection unit), 14 is an image processing unit (two-dimensional blood vessel image correction unit, enlargement / reduction unit), and 15 is 2. A dimension feature point extraction unit (two-dimensional feature point extraction unit), 16 a change detection unit (change detection unit, distortion detection unit), 18 a two-dimensional matching feature point generation unit (two-dimensional matching feature point generation unit), Reference numeral 19 is a collation unit (collation means), and 31 is a three-dimensional feature point database (three-dimensional feature point storage means).

Claims (5)

A two-dimensional image of a blood vessel in a living body is acquired to obtain a two-dimensional blood vessel image, and at least three points belonging to the object to be imaged at the time of imaging are used as reference points, and distances from a predetermined position of the biometric identification device to each reference point are determined. a photographing means for detecting the inclination of the distance and the reference plane to the reference plane from a predetermined position of the measurement to create and biometric device reference surface to be imaged at the time of shooting,
Two-dimensional feature point extracting means for analyzing the two-dimensional blood vessel image acquired by the photographing means and extracting two-dimensional feature points of the blood vessel pattern;
The posture of the living body at the time when the blood vessel image that is the basis of the three-dimensional feature point of the blood vessel pattern is acquired is set as the basic posture, and is detected by the imaging unit when the two-dimensional blood vessel image is acquired by the imaging unit. It has been the attitude of the corresponding two-dimensional blood vessel image to the inclination of the distance and the reference plane from a predetermined position to the reference plane of the biometric identification device by the verification during posture, detecting the degree of change to the verification during orientation from the basic position Change detection means;
Corresponds to the distance from the predetermined position of the biometric identification device detected by the photographing means to the reference plane and the inclination of the reference plane, and the degree of change from the basic attitude detected by the change detecting means to the matching attitude. A three-dimensional feature point of the blood vessel pattern when the living body is in the basic posture is acquired from the three-dimensional feature point storage means, and the change detection is performed from the three-dimensional feature points of the acquired blood vessel pattern. Two-dimensional matching feature point generating means for generating a two-dimensional matching feature point of the blood vessel pattern by plotting a three-dimensional feature point corresponding to the degree of change detected by the means on the clipping plane;
A matching means for matching the two-dimensional matching feature point of the blood vessel pattern generated by the two-dimensional matching feature point generating means with the two-dimensional feature point of the blood vessel pattern extracted by the two-dimensional feature point extracting means; A biometric identification device comprising: The biometric identification device according to claim 1,
Comprising strain detection means for detecting the degree of distortion from the basic posture to the posture at the time of matching;
The biometric identification device characterized in that the two-dimensional matching feature point generating means generates a two-dimensional matching feature point of a blood vessel pattern corresponding to the degree of distortion detected by the distortion detecting means. The biometric identification device according to claim 1,
Distortion detecting means for detecting the degree of distortion from the basic attitude to the matching attitude;
Two-dimensional blood vessel image correction means for correcting the two-dimensional blood vessel image acquired by the imaging means in accordance with the degree of distortion detected by the distortion detection means;
The biometric identification device, wherein the two-dimensional feature point extraction unit analyzes the two-dimensional blood vessel image corrected by the two-dimensional blood vessel image correction unit and extracts a two-dimensional feature point of a blood vessel pattern. In the biometric identification device according to any one of claims 1 to 3,
An outer shape detecting means for detecting the outer shape of the living body;
Enlarging / reducing the two-dimensional blood vessel image acquired by the imaging unit based on the detection result of the outer shape detecting unit to enlarge / reduce the size of the blood vessel image that is the basis of the three-dimensional feature point of the blood vessel pattern. Means and
The biometric identification device, wherein the two-dimensional feature point extracting unit extracts a two-dimensional feature point of a blood vessel pattern by analyzing the two-dimensional blood vessel image enlarged or reduced by the enlargement / reduction unit. The biometric identification device according to any one of claims 1 to 4,
The biometric identification apparatus characterized in that the two-dimensional matching feature point generating means generates three-dimensionally each of the two-dimensional matching feature points of the blood vessel pattern.

Images