JPH1153545A - Device and method for collation - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform collation through a collating method corresponding to the condition of a reagent. SOLUTION: The image of the fingerprint of a finger inputted from an input sensor 61 is displayed on a display device 72, supplied to a binarizing processing part 62 and binarized based on a threshold set by a threshold setting part 66 and after smoothing processing is performed at a smoothing processing part 63, the direction of a ridge line is detected by a ridge line direction detecting part 64. Ridge line direction angle data showing the ridge line direction are generated at a ridge line direction angle data forming part 65, and these data are collated with ridge line direction angle data, which are previously stored in a registered ridge line direction angle data storage part 68, corresponding to the said finger. At such a time, an area having disturbance such as an injury inputted from a key input part 69 and set by a collation area setting part 67 is excluded from the object of collation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a collating apparatus and method, for example, a collating apparatus and method for reading a fingerprint of a human fingertip and collating a fingerprint image generated based on the fingerprint with a comparative reference fingerprint image. About.

[0002]

2. Description of the Related Art Confirming the identity of a person is one of important information security techniques. Among them, research and development of personal identification methods such as fingerprint collation and palm print collation using individual physical characteristics are active. In particular, fingerprints and palm prints are considered to have extremely excellent characteristics such as "everyone is unidentified" and "lifelong invariant", and their commercialization as well as R & D is active. For example, in the case of fingerprint collation, fingerprint data of each person is registered in advance, and fingerprint data of a sample (finger) detected by a fingerprint sensor is collated with the registered data.

As shown in FIG. 13, there is a fingerprint sensor using, for example, total internal reflection of the inner surface of a right-angle prism. The fingerprint sensor irradiates irradiation light from a light source such as the light emitting diode 2 to the inner surface 3A of the predetermined reflecting surface of the right-angle prism 3 at an angle that satisfies the condition of total reflection. At this time, a fingertip (sample) F to be detected as a fingerprint is pressed against the outside of a predetermined reflecting surface of the right-angle prism 3, and the reflected light L 2 from the reflecting surface is imaged by the imaging unit via the imaging optical system 4. An image is formed on a CCD (charge coupled device) 5 and the reflected light L2 is photoelectrically converted to read a fingerprint image as photoelectric information.

When actually performing fingerprint collation, a fingerprint image is generated based on photoelectric information based on a reflected light pattern PR obtained by photoelectric conversion, and this fingerprint image is compared with a fingerprint image which is a reference registered in advance. Matching is performed using a technique such as pattern matching.

In this case, the fingerprint image is detected when the sample is pressed against the reflecting surface of the right-angle prism 3 because the refractive index of the contact portion is different from that of the non-contact portion due to air. The detection principle is used in which the reflection condition is broken and a reflected light pattern PR corresponding to the contact area is obtained.

[0006]

However, fingerprints and palm prints may temporarily change due to cuts, abrasions, burns, etc. In this case, there is a problem that fingerprint collation may not be performed normally. .

[0007] The present invention has been made in view of such a situation, and it is an object of the present invention to enable normal fingerprint collation even if the fingerprint changes due to cuts, abrasions, burns and the like.

[0008]

According to a first aspect of the present invention, there is provided a collating apparatus for collating input means for inputting an image of a sample, an image input by the input means, and a previously registered image of the sample. It is characterized by comprising a collating means and a switching means for switching the collating method of the collating means according to the state of the sample.

According to a third aspect of the present invention, there is provided a collating apparatus comprising: input means for inputting an image of a sample; conversion means for converting the image input by the input means into digital data based on a predetermined reference value; Instructing means for instructing the situation, changing means for changing the reference value according to the situation of the specimen instructed by the instructing means, and detecting means for detecting the characteristic amount of the specimen from the digital data converted by the converting means ,
It is characterized in that it comprises a collation unit for collating data representing the characteristic amount of the specimen detected by the detection unit with data representing the characteristic amount of the specimen registered in advance.

[0010] According to a fifth aspect of the present invention, there is provided a collating apparatus comprising: an input unit for inputting an image of a specimen; a specifying unit for specifying a first area where disturbance has occurred in an image input by the input unit; In the image input by the input unit, the second region excluding the first region specified by the specifying unit is compared with a region corresponding to the second region of the sample image registered in advance. Collating means.

According to a sixth aspect of the present invention, there is provided a collating method comprising: an inputting step of inputting an image of a specimen; a collating step of collating the image inputted in the inputting step with an image of the specimen registered in advance; And a switching step of switching the collation method according to the state of the sample.

[0012] According to a seventh aspect of the present invention, there is provided a collating method comprising the steps of: inputting an image of a sample; converting the image input by the input step into digital data based on a predetermined reference value; An instruction step of instructing a situation, a change step of changing a reference value according to the situation of the specimen instructed by the instruction step, and a detection step of detecting a characteristic amount of the specimen from digital data converted by the conversion step. And a collation step of collating data representing the characteristic amount of the specimen detected in the detection step with data representing the characteristic amount of the specimen registered in advance.

[0013] In the collation method according to the present invention, an input step of inputting an image of the sample, a specifying step of specifying a first area where disturbance occurs in the image input in the input step, In the image input in the input step, the second area excluding the first area specified in the specifying step is compared with an area corresponding to the second area of the image of the specimen registered in advance. And a collation step.

A collation device according to claim 1 and claim 6.
In the matching method described in the above, the input image of the sample,
When collating with a pre-registered image of the specimen, the collation method is switched according to the state of the specimen.

A collation device according to claim 3 and a collation device according to claim 7.
In the matching method described in the above, when converting the image of the input sample into digital data based on a predetermined reference value, the reference value is changed according to the status of the specified sample, and based on the reference value The characteristic amount of the specimen is detected from the converted digital data, and data representing the characteristic amount of the detected specimen is collated with data representing the characteristic amount of the specimen registered in advance.

A collation device according to claim 5 and an collation device according to claim 8
In the collation method described in 1 above, in the input image of the specimen, the first area where disturbance has occurred is specified, and the second area excluding the specified first area and the second area registered in advance are registered. The region corresponding to the second region of the image of the sample is collated.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a block diagram showing an example of the configuration of an embodiment of a fingerprint matching device to which the matching device of the present invention is applied. The detection unit 11 includes an optical sensor 12 and a pressure sensor 13, and reads information on a sample. The optical sensor 12 is configured to capture the reflected light of the irradiation light applied to the specimen by an imaging unit including a photoelectric conversion element and the like, convert the reflected light into photoelectric information S1, and then supply the photoelectric information S1 to the image processing unit 14. I have.

The image processing section 14 performs grayscale image processing and binarization processing on the photoelectric information S1 based on the reflected light pattern from the sample, and then compares the image data S3 generated by the binarization processing with data. The data is sent to the unit 15.

On the other hand, the pressure sensor 13 is an optical sensor 1
2, when the photoelectric information S1 is detected, the pressing force of the sample against the detecting unit 11 is detected and sent to the data matching unit 15 as pressure data S2. Thereby, the data collating unit 15 can receive the pressure data S2 and the image data S3 as a set data.

The data collating unit 15 reads the image data of the comparison reference with a plurality of different pressing forces in advance, and sets the image data and the set data of the pressing force data at that time as a set of comparison reference data for each sample F in a database. 15A
Registered within.

When the data collating unit 15 receives the image data S3 of the fingerprint of the sample F read by a predetermined pressing force via the image processing unit 14, it matches the pressure data S2 of the set data of the image data S3. The pressing force data of the set data of the comparison standard of the sample F is searched. Then, the image data S3 is collated using the image data of the set data of the pressing force data.

As shown in FIG. 2, the optical sensor 12 is formed by cutting a fiber optic plate (FOP) 20 in which a plurality of linear optical fibers are bundled obliquely at a predetermined angle from a side surface. An optical surface is formed from the surface, and this optical surface is used as a detection surface 20A for detecting image data of the sample F. And this detection surface 20A
The sample F is pressed against the outer surface of the light-emitting diode 20, and the light-emitting diode 21 as a light source is pressed against the inner surface of the detection surface 20 </ b> A.
The irradiation light L1 is irradiated at an incident angle satisfying the total reflection condition from 0B. Then, the reflected light L2 from the inner surface of the detection surface 20A
Is directly incident on the CCD 22 connected to one end 20C of the FOP 20 via the FOP 20.

The CCD 22 makes the reflected light L2 of the sample F at the contact portion with the detection surface 20A incident on the light receiving surface, photoelectrically converts this into photoelectric information S1, and outputs it to the image processing unit 14.

A pressure sensor 13 composed of, for example, a piezo-type piezoelectric element is disposed at four predetermined positions corresponding to positions at which the detection portion of the sample F is pressed against the surface of the detection surface 20A of the FOP 20. The pressure sensor 13 converts the pressing force when the sample F is pressed against the detection unit 11 into a distortion voltage and outputs the distortion voltage to the data collation unit 15 as pressure data of an electric signal.

Also, by disposing the pressure sensors 13 at the four positions on the detection surface 20A, each pressure sensor 1
3, each pressure sensor 1 based on the pressure data S2 detected from
The sample F is pressed so that pressure is applied evenly to the sample 3, and the sample F can be pressed against the detection surface 20A in an optimal contact state suitable for detecting surface information. Thus, fingerprint image data can be read with high accuracy.

Actually, as shown in FIGS. 3A and 3B, the fingerprint FP1 (FIG. 3A) when the sample F is lightly pressed against the detection surface 20A and the sample F Detection surface 2
Fingerprint FP2 when strongly pressed against 0A (FIG. 3)
(B)), ridges (convex portions) of the fingerprint FP1
The contact area of the ridge of the fingerprint FP2 increases as compared with the contact area of the fingerprint FP2. As a result, the reflected light patterns P1 and P2 of the fingerprints FP1 and FP2 are read as different.

As described above, even when the fingerprint of the same sample F is read, the image data obtained by the pressing force of the sample F against the detection surface 20A is different.

Therefore, in the data collating unit 15, 1
For each of the samples F, the image data of the comparison reference is read for each of a plurality of different pressing forces, and the pressing force data and the image data at that time are registered in the database 15A as a set data of the comparison reference. Then, when actually performing fingerprint collation, the data collation unit 15 first reads from the database 15A a plurality of sets of comparison criteria relating to the sample F pressed against the detection surface 20A, and reads the pressure data S2 from the plurality of sets of data. Search for the set data having the pressing force data matching with. Then, of the pressing force data of the plurality of sets of data, one of the pressing force data is the pressure data S.
When it is determined that the set value matches 2, the image data S3 is compared with the reference image data of the set data having the pressing force data.

Thus, when the specimen F is brought into contact with the detection surface 20A with various different pressing forces and collation is performed, the pressing force data of the comparison standard is read from the database 15A according to the pressure data S2 at that time. Accordingly, the image data can be collated using the image data read under the same conditions of the pressing force. In this manner, based on the pressure data S2, it is possible to perform reliable image data matching according to the deformation of the fingerprint shape of the sample F.

As a method of comparing a fingerprint image in the data matching unit 15, direction angle data obtained by encoding the symmetric flow of fingerprint ridges is obtained, and direction code matching or collation is performed based on the direction angle data. A feature point matching method using feature point data such as an end point or a branch point of a ridge of a fingerprint is used.

In the above configuration, the sample F is
Of the detection surface 20A.
Irradiation light L1 is emitted from the light emitting diode 21 to the inner surface of A at an incident angle that satisfies the condition of total reflection, and reflected light L2 from the inner surface of the detection surface 20A is imaged by the CCD 22.
As a result, information such as the shape or pattern of the surface of the sample pressed on the detection surface 20A is used as a reflected light pattern as a CC pattern.
It can be read by D22.

The CCD 22 converts the reflected light L2 into photoelectric information S1.
And sends it to the image processing unit 14.

The image processing section 14 can generate image data S3 corresponding to the reflected light pattern of the sample F by performing predetermined image processing based on the photoelectric information S1.

At this time, the detecting unit 11 simultaneously
The pressing force when the sample F is pressed against the detection surface 20A is read by the pressure sensor 13 disposed on the surface of the detection unit 20A as pressure data S2, and is sent to the data collation unit 15.

The data collating unit 15 includes a database 15A
The data having the pressing force data that matches the pressure data S2 is searched from the set data of the comparison criterion relating to the sample F registered therein. Then, when there is pressing force data that matches the pressure data S2 from the registered comparison reference set data, the data matching unit 15 reads the pressing force data set data as comparison reference image data. Collated with the image data S3.

As a result, in the data collating unit 15,
The pressing force of the sample F on the detection surface 20A is the same predetermined pressing force as when the image data of the comparison standard is read,
The image data S3 at that time can be compared with the image data of the comparison reference.

Therefore, even when the shape of the sample F is deformed, the image data S3 can be compared with the comparison data by reproducing and collating almost the same shape deformation caused by the same pressing force on the sample F. Can be compared with the image data with high accuracy, and the reliability of crime prevention can be improved.

According to the above configuration, the pressure data S2 of the pressing force of the sample F against the detection surface 20A is read by the pressure sensor 13 in the detection unit 11, and the data collation unit 1
5, the image data S3 is collated using the image data of the comparison reference read with the pressure data that matches the pressure data S2. Thus, even when the shape of the sample F is deformed by the pressing force, it is possible to perform highly accurate fingerprint collation using image data corresponding to the deformation.

Further, when the collation is performed using the copy image of the fingerprint, the image data of the copy image is not the pressing force at the time of reading this image data, that is, both the pressure data S2 and the image data S3 are pressed. If the pressure data does not match the image data, the data collation unit 15 determines that the collation is inappropriate, so that the reliability of crime prevention can be improved.

In the first embodiment,
A case has been described in which image data with a plurality of different pressing forces are registered in the database 15A as set data of a comparison reference, and the sample F is collated under one of the pressing force conditions. Without being limited thereto, based on a plurality of different reference data sets for one sample F, fingerprint image data is read a plurality of times with different pressing forces corresponding to the respective reference data sets, The image data S3 for each pressing force may be compared with the image data of the comparison reference. In this case, if it is determined that the collation is appropriate only when all of the image data S3 read by a plurality of different pressing forces match the corresponding comparison reference image data, the deformation of the sample F due to the pressing force is determined by a plurality of different pressing forces. Collation reproduced for each pressure can be performed. Thereby, the accuracy of fingerprint collation can be further improved.

Further, the data collating unit 15 stores a sufficient amount of pressure data and a set of image data corresponding thereto so that the relationship of the deformation of the image data according to the pressure of the sample F can be inferred. And by programming the analogy method, even if the pressure data S2 of the pressing force of the sample F does not completely match the registered pressing force data,
The image data may be analogized by inferring the image data from the pressing force data close to the pressure data S2. In short, the pressing force data of a predetermined comparison reference is set based on the pressure data S2 when the sample F is pressed. Then, the image data S3 may be compared with the image data of the comparison reference corresponding to the pressing force data.

FIG. 4 is a block diagram showing a configuration example of a second embodiment of a fingerprint matching device to which the matching device of the present invention is applied. In the fingerprint matching device shown in FIG. 4, a display unit 26 for displaying pressure data S2 input from the pressure sensor 13 is newly provided in the fingerprint matching device shown in FIG. Other configurations and operations are shown in FIG.
, The corresponding parts are denoted by the same reference numerals, and description thereof will be omitted as appropriate.

In the fingerprint collation apparatus shown in FIG. 4, when the sample F comes into contact with the predetermined detection surface 20A of the optical sensor 12 of the detection unit 11, the pressure data S output from the pressure sensor 13
2 is read and the read numerical data of the pressure data S2 is displayed on the display unit 26 to notify the subject.

Further, in the fingerprint collation apparatus shown in FIG. 4, when registering data as a comparison reference for collation,
First, the pressing force pressed against the detection surface 20A of the sample F is read by the pressure sensor 13 and displayed on the display unit 26 to notify the registrant, and the image data and the pressing force data at that time are displayed on the sample F. Is registered in the database 15A as a set of predetermined comparison criteria.

Here, when fingerprint matching is performed by the data matching unit 15, first, the pressing force of the sample F on the detection surface 20A is displayed on the display unit 26. In response to this, the examinee refers to the numerical data displayed on the display unit 26 and, when the predetermined pressing force at the time of registration known only by the registrant of the comparison reference data is reached, the subject Press the input switch (not shown). In response, the data matching unit 15 reads, via the optical sensor 12, the image data S3 of the sample F when the input switch is pressed, and checks the image data S3 with the image data of the reference for the sample F.

As a result, the registrant of the data sets the detection surface 20
The sample F can be pressed against A with the same pressing force as the predetermined pressing force when the image data of the comparison reference is read. Thereby, the collation can be performed using the image data S3 with the same pressing force as the image data of the comparison reference. In this manner, the collation using the image data according to the pressing force of the sample F on the detection surface 20A can be performed.

At this time, the numerical data of the predetermined pressing force of the image data of the comparison reference is made known only to the registrant. The pressure data S2 to be displayed on the display unit 26
For the numerical data of, an arbitrary unit may be set.

Accordingly, if the actual data of the predetermined pressing force when the registrant registers the image data of the comparison reference is not known, for example, the image data of the fingerprint of the registrant's sample F is copied. Is duplicated, the copied image data is based on the predetermined pressing force and is not pressed by the predetermined pressing force at the time of the verification, and the data verification unit 15 determines that the comparison data is compatible. Can not get. As a result, it is possible to prevent the matching of the collation using the copy of the fingerprint beforehand, and it is possible to improve the reliability of crime prevention.

In the above configuration, similarly to the first embodiment, the reflected light L2 of the specimen F pressed against a predetermined position on the detection surface 20A is received by the CCD 22, and the surface by the reflected light L2 After the shape information is photoelectrically converted into photoelectric information S1 by the optical sensor 12, the information is sent to the image processing unit 14. Then, the image processing unit 14 generates fingerprint image data S3.

At this time, the detecting unit 11 simultaneously
The pressing force when the sample F is pressed against the detection surface 20A is read as pressure data S2 by the pressure sensor 13 disposed on the surface of the detection surface 20A, and the data collation unit 15
Sent to

Then, based on the pressure data S2, the data collating unit 15 displays the numerical data on the display unit 26 to notify the subject. Thus, the subject can determine from the numerical data of the pressure data S2 displayed on the display unit 26 whether or not the pressing force on the detection surface 20A of the sample F has reached a predetermined pressing force known only to the registrant of the data. Can be determined. Then, the subject A presses the input switch when the pressing force displayed on the display unit 26 reaches a predetermined numerical value. Thus, the data collating unit 15 performs fingerprint collation with the comparison reference image data using the image data S3 read by the optical sensor 12 at this time.

Therefore, according to the above configuration, in the data collating unit 15, the pressing force of the sample F on the detection surface 20A is set to the same predetermined pressing force as when the comparative reference image data is read. Image data can be collated.

Therefore, by reproducing almost the same deformation of the sample F due to the same pressing force, it is possible to perform highly accurate collation between the image data S3 and the image data of the comparison standard, and to improve the reliability of crime prevention. Can be improved.

Further, only the registrant of the data can know the pressing force at the time of pressing the input switch. If the pressing force is different from that at the time of registering the reference image data, the collation is inappropriate. By doing so, highly accurate collation can be performed, and the reliability of crime prevention can be further improved.

Note that in the second embodiment,
Although the description has been given of the case where the pressure data S2 by the pressing force of the sample F on the detection surface 20A is displayed as numerical data on the display unit 26, the present invention is not limited to this, and the display may be stepwise displayed by light emitting diodes or the like. It is possible. Alternatively, the notification may be made by a sound such as a buzzer.

Further, the pressing force data of the set data registered in the database 15A is set to a predetermined pressing force in advance, and the pressure data S2 of the sample F read by the data collating unit 15 and the predetermined pressing force May be compared, and the result of the comparison may be displayed on the display unit 26, thereby instructing the subject to press and adjust the sample F so that the pressing force of the sample F becomes a predetermined value. Thus, the fingerprint image data S3 can be read with the same predetermined pressing force suitable for image matching as when the comparison reference image data is read. Accordingly, it is possible to prevent the fingerprint of the subject from being missed.

In the above embodiment, the sample F
Although the case where the pressing force against the detection surface 20A of the above is measured by the pressure element has been described, the present invention is not limited to this.
For example, as shown in FIG. 5, a guide 31 formed of transparent silicon rubber or the like is provided on an imaging device (not shown),
Strain gauges 32 are provided in the right and left regions of the concave portion 31A of the guide 31 to detect pressure strains occurring in the surface strain of the sample F, respectively, and to use the pressure strains as a set data with the image data. Is also good.

Thus, the image data of the sample F detected by the guide 31 can be determined in accordance with the pressure distortion, and the fingerprint-shaped pressure distortion generated when the sample F is pressed against the detection surface 20A. Fingerprint matching corresponding to the deformation due to.

In this case, by disposing the strain gauges 32 in the right and left regions of the concave portion 31A, the specimen F
Is detected by the strain gauge 32 in the pressing direction to the right area or the left area when the object is pressed against the detection surface 20A, and it is determined whether the pressing direction is appropriate according to the bias. Then, the bias in the pressing direction may be corrected.

Further, by registering image data corresponding to the bias of the predicted pressing direction in the database 15A, the sample F can be detected on the detection surface 20A based on the image data stored in the database 15A. It is also possible to realize image data collation independent of the direction in which the image data is pressed.

Further, for example, if the uneven shape of the sample F is stored in a shape memory material, pressure data is obtained for each part of the uneven shape, and the pressure data is used as a set data with the image data. Matching with high accuracy can be performed.

FIG. 6 is a block diagram showing a configuration example of a third embodiment of a fingerprint matching device to which the matching device of the present invention is applied. The detecting unit 51 is configured, for example, as shown in FIG. It has been made like that. At this time, when a fingertip (sample) as a fingerprint detection target is pressed against the outside of the predetermined reflecting surface of the right-angle prism 3, the reflected light L <b> 2 from the reflecting surface forms on the CCD 5 via the imaging optical system 4. Imaged. The CCD 5 photoelectrically converts the reflected light L2 and reads a fingerprint image as photoelectric information.
This photoelectric information corresponds to a gray image corresponding to the unevenness of the fingertip.

After binarizing the photoelectric information, the image processing section 52 smoothes the image by a spatial filter that takes a majority decision of 9 pixels including 8 peripheral pixels in order to eliminate the influence of noise such as sweat pores. Has been made. The fingerprint collation system 53 generates fingerprint data by extracting the characteristic amount of the photoelectric information from the image processing unit 52, and compares the fingerprint data with the same reference sample registered in the registered fingerprint database storage unit 54 in advance. Is collated with fingerprint data by using a technique such as pattern matching.

Next, the operation will be described. From the light emitting diode 2 to the inside 3 of a predetermined reflecting surface of the right-angle prism 3
Irradiation light is applied to A at an angle that satisfies the total reflection condition. At this time, when a fingertip (sample) F as a fingerprint detection target is pressed against the outside of the predetermined reflecting surface of the right-angle prism 3, the reflected light L 2 from the reflecting surface is applied to the CCD 5 via the imaging optical system 4. It is imaged. This reflected light L2 is CC
The photoelectric conversion is performed by D5, and the fingerprint image is read as photoelectric information. This photoelectric information corresponds to the grayscale image corresponding to the unevenness of the fingertip as described above.

This photoelectric information is supplied to the image processing section 52, and after being binarized, is subjected to a spatial filter that takes a majority decision of nine pixels including eight peripheral pixels in order to remove the influence of noise such as sweat pores. Smoothed. After that, the fingerprint data is supplied to the fingerprint collation system 53, the feature amount thereof is extracted, and pattern matching (collation) is performed with the fingerprint data serving as a comparison reference of the sample F registered in the registered fingerprint data storage unit 54.

As a method of extracting a feature value, for example,
A feature amount is extracted from a point called a minutia, as shown in FIG. 7, where the ridge of the fingerprint is interrupted (end point), or as shown in FIG. 8, a point at which the ridge of the fingerprint branches (branch point). As shown in FIG. 9, the fingerprint image is divided into predetermined blocks (for example, 16 pixels × 16 pixels), the direction of the ridge of the fingerprint in each block is determined, and the flow direction of the ridge ( There is a method of extracting the characteristic amount of the fingerprint from the ridge direction), a method of extracting the characteristic amount from the pattern of the entire fingerprint, and the like. In some cases.

For example, when extracting the characteristic amount of the fingerprint from the ridge direction, the input fingerprint data and the registered fingerprint data registered in the registered fingerprint data storage unit 54 are converted into this format, and the ridge direction angle data is extracted. If the degree of coincidence (in this example, the number of blocks in which the ridge directions coincide among the blocks shown in FIG. 9) is greater than or equal to a predetermined threshold value, the input fingerprint data and the registered fingerprint data Storage unit 54
Is determined to be the same as the registered fingerprint data, and the identity of the user is confirmed (verification is OK).

In this case, if there are cuts, abrasions, hand marks, burns, etc., these appear as ridge direction angle data,
Since it becomes a disturbance, it becomes a large barrier at the time of matching. Therefore, it is necessary to be able to separate this from the fingerprint on the system.

FIG. 10 is a block diagram showing a detailed example of the configuration of a fingerprint matching device which detects a feature amount of a fingerprint from the ridge direction and performs matching. The input sensor 61 (input means) corresponds to the detection unit 51 in FIG. 6 and has a configuration as shown in FIG. 13, for example.
An image is captured by D5 and a corresponding signal is output. Key input unit 69 (switching means, instruction means)
Is constituted by a keyboard, a mouse, or the like, and is configured to input a predetermined key indicating that there is an abrasion, a hand touch, a cut or a burn, or an area to be excluded from binarization. The display device 72 displays an image obtained by the input sensor 61.

The threshold setting section 66 (change means) sets a threshold for binarization used in the binarization processing section 62, and the area setting section 71 sets the threshold setting section 66
Is to set each area when setting a threshold value for each of a plurality of areas obtained by dividing the screen. The binarization processing section 62 (conversion means) includes an input sensor 61
The data corresponding to the fingerprint image is binarized based on the threshold value set for each area set by the area setting unit 71 supplied from the threshold setting unit 66. I have.

The smoothing section 63 performs a smoothing process on the binarized data. The ridge direction detection unit 64 (detection means) determines the ridge direction of the data smoothed by the smoothing processing unit 63 as 16 pixels × 1.
The detection is performed for each block of 6 pixels. The ridge direction angle data forming unit 65 is configured to generate ridge direction angle data as shown in FIG. 9 from the ridge direction detected by the ridge direction detection unit 64.

The collation area setting section 67 (specification means) sets an area to be collated based on the area input by the key input section 69. The registered ridge direction angle data storage unit 68 stores ridge direction angle data of a fingerprint to be collated in advance. The matching processing unit 70 (matching means) includes a ridge direction angle data forming unit 65
The matching ridge direction angle data corresponding to the fingerprint image input by the input sensor 61 and supplied from the matching area set by the matching area setting unit 67 are stored in the registered ridge direction angle data storage unit 68. The corresponding ridge direction angle data is compared with the data of the portion corresponding to the matching area, and the matching result is output.

Next, the operation will be described. When a finger is pressed against the prism surface of the input sensor 61 and a predetermined light is applied to the prism surface, a grayscale image corresponding to the degree of unevenness of the fingerprint of the finger is obtained, and data corresponding to the grayscale image of the fingerprint is 2. It is supplied to the value processing section 62 and the display device 72. Data corresponding to the fingerprint image of the sample (finger) supplied from the input sensor 61 is supplied to the display device 72,
The fingerprint image is displayed and supplied to the binarization processing unit 62. The user informs the fingerprint collation device of the fingerprint image displayed on the display device 72 by inputting a corresponding key indicating that there is a cut, a burn, an abrasion, a hand fray, and the like, as shown in FIG. Specify the area. Data indicating the designated area is supplied to the collation area setting unit 67, and an area where collation is not performed is set.

The fingerprint image data supplied to the binarization processing unit 62 is set for each predetermined block (for example, a block of 4 pixels × 4 pixels) supplied from the threshold value setting unit 66. Binarization is performed based on the threshold value. For example, the threshold value is set relatively high in an area having a scratch or the like. If the fingerprint is thin due to hand rubbing or the like, the threshold value is set relatively low in that area.

The fingerprint image data binarized based on the threshold value set by the threshold value setting unit 66 is supplied to the smoothing processing unit 63. The smoothing processing unit 63 removes an image (noise) such as a sweat pore included in the image data of the fingerprint by, for example, a spatial filter that takes a majority decision using nine pixels including eight pixels around a predetermined target pixel. Performs smoothing. Next, the feature amount of the smoothed image data is extracted by the ridge direction detection unit 64. For example, 16 pixels × 16 pixels of smoothed fingerprint image data
A global flow of ridges is detected for each pixel. And
The ridge direction angle data forming unit 65 generates ridge direction angle data obtained by encoding the directional angle of the ridge, and supplies the ridge direction angle data to the matching processing unit 70.

The ridge direction angle data supplied to the matching processing unit 70 is compared with the corresponding ridge direction angle data stored in the registered ridge direction angle data storage unit 68. At this time, based on the data indicating the area not to be collated supplied from the collation area setting unit 67, the collation is performed on the area excluding the area not collated. When matching is performed based on the ridge direction, the data stored in the registered ridge direction angle data storage unit 68 indicates the global flow of the ridge every 16 pixels × 16 pixels of the input fingerprint image data. The ridge direction angle data shown should be in the format.

The matching processing unit 70 compares the ridge direction angle data supplied from the ridge direction angle data forming unit 65 with the ridge direction angle data stored in the registered ridge direction angle data storage unit 68. Then, it is determined whether or not the user is a person based on whether or not the degree of matching is equal to or more than a predetermined reference value. That is, if the degree of those that match is greater than or equal to a predetermined reference value, it is determined that the person is the person, and if the degree of those that match is less than the predetermined reference value,
It is determined that he is not himself. Then, a signal corresponding to the determination result is output.

FIG. 12 shows another example of a method for designating a predetermined area of a finger. In the case of this example, the display device 72
Are displayed on the screen, and numbers are assigned to respective areas obtained by dividing the finger figure. Therefore, the user can specify an area corresponding to the selected number by selecting one or more numbers. The area specified in this way is set as an area where the matching process is not performed in the matching area setting unit 67.

In addition, in the case of a linear flaw, an area to be excluded can be specified by designating the start and end thereof on the image displayed on the display device 72. It is. Then, the designated area is excluded from the matching processing.

In addition, with respect to disturbances that exist in a wide range and are thin, such as scratches and hand rubbing, the threshold value of binarization can be increased to prevent the scratches from being recognized as fingerprints. On the other hand, by lowering the threshold value for binarization, it is possible to extract a fingerprint that is likely to disappear due to hand shaking. In this case, increase the threshold,
The area to be lowered can be input from the key input unit 69.

For a disturbance such as a cut or a burn that remains as a clear image, the area designated by the key input unit 69 is excluded from the target of the collation processing when the collation processing is performed. In the case where the ridge direction is detected and collated, the ridge direction may be detected for an area which is previously excluded from the area designated by the key input unit 69.

Further, if the collation processing unit 70 does not set the collation area from the key input unit 69, and if only the local part has a low degree of coincidence and only linear non-coincidence, the collation is OK. It is also possible to do so. Also in this case, it is possible to use a mode switch or the like to indicate that there is a disturbance such as a cut. In addition, when there is a thin and wide disturbance such as a scratch, the difference amount between the input image and the registered image is calculated for each area, and the difference amount in each area is equal to or less than a predetermined amount. , Collation OK.

As described above, the disturbance such as a scratch can be processed separately from the fingerprint or palm print, so that a cut may be made with a kitchen knife or the like, thereby causing a problem that the person cannot be authenticated as a person. Can be suppressed, and the reliability of the matching process can be improved. In addition to being able to identify areas where disturbances such as wounds are present, and by actively disclosing the person's physical condition (for example, cuts and abrasions), the person's physical condition Is performed, and even if the physical condition of the person changes, the person acceptance rate at which the person is authenticated as the person can be improved.

In the first and second embodiments, the optical sensor 12 has a detection surface 20 for the specimen F.
Although the case where A is provided on the end surface of the FOP 20 has been described, an optical prism may be used instead of the FOP 20 and a predetermined reflection surface of the optical prism may be used as the detection surface. Further, the image of the sample F pressed on the detection surface may be directly read by a solid-state imaging device such as a CCD.

In each of the above embodiments, the case where the light emitting diode is used as the light source of the optical sensor 12 for optically detecting the fingerprint shape has been described. However, the present invention is not limited to this. A light source such as a device may be used.

In the above embodiment, the case where the reflected light is converted into photoelectric information, that is, an electric signal by the photoelectric conversion element to create a fingerprint image has been described. However, the present invention is not limited to this. The reflected light pattern, which is information,
An image captured by a camera or the like may be directly taken as photographic image information, and image matching may be performed based on the photographic image information.

Further, in the above-described embodiment, a case has been described where the fingerprint detection device detects a fingerprint of a human fingertip. However, the present invention is not limited to this. A palm print or a fingerprint of a toe may be used. Further, the present invention can be applied to an image other than a person's identification such as a cow's nose, which changes an image read when pressed against a detection surface.

The information to be read is not limited to the fingerprint of a person or an animal, but may be any as long as the image data to be read changes according to the pressing force on the detection surface. be able to.

In the above-described embodiment, the key input unit 69 is used to input a key indicating the state of the finger or palm such as abrasion or hand sparing. It is also possible to switch the mode depending on whether or not there is a hand. Then, the threshold value setting section 66 can change the threshold value according to the mode set by the mode changeover switch.

When the area is not designated by the key input section 69, the display device 72 can be omitted in FIG.

Further, in the third embodiment, the input sensor 61 is constituted by a prism as shown in FIG. 12, but is constituted by an FOP as shown in FIG. It is also possible.

[0092]

According to the collation apparatus according to the first aspect and the collation method according to the sixth aspect, the image of the input specimen is
When collating with a pre-registered sample image, the collation method is switched according to the condition of the sample,
The collation can be performed by an optimal collation method according to the state of the sample, and the reliability of the collation can be improved.

The collating apparatus according to claim 3 and claim 7
According to the matching method described in the above, when converting the image of the input sample into digital data based on a predetermined reference value,
The reference value is changed according to the condition of the specified sample, and the feature amount of the sample is detected from the digital data converted based on the reference value, and the data representing the feature amount of the detected sample is registered in advance. Since the data is collated with the data representing the characteristic amount of the sample, optimal digitization can be performed according to the condition of the sample, and the reliability of the comparison can be improved.

The collating apparatus according to claim 5 and claim 8
According to the collation method described in 1 above, of the input sample image, the first area where disturbance has occurred is specified, and the second area excluding the specified first area is registered in advance. Since the region corresponding to the second region of the image of the sample is collated, even if the condition of the sample changes, the collation can be performed using the unchanged portion, and the reliability of the collation can be improved. Performance can be improved.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration example of a first embodiment of a fingerprint matching device to which the present invention is applied.

FIG. 2 is a diagram illustrating a configuration example of a detection unit 11;

FIG. 3 is a diagram for explaining a change in a reflected light pattern due to a difference in pressing force.

FIG. 4 is a block diagram illustrating a configuration example of a second embodiment of a fingerprint matching device to which the present invention is applied;

FIG. 5 is a diagram illustrating a configuration example of a detection unit 11 provided with a strain gauge.

FIG. 6 is a block diagram illustrating a configuration example of a third embodiment of a fingerprint matching device to which the present invention has been applied.

FIG. 7 is a diagram showing end points of a fingerprint line.

FIG. 8 is a diagram showing a branch point of a fingerprint line.

FIG. 9 is a diagram showing ridge direction angle data.

FIG. 10 is a block diagram illustrating a detailed configuration example of a fingerprint matching device.

FIG. 11 is a diagram illustrating a method of designating a damaged area.

FIG. 12 is a diagram illustrating a method of specifying an area using numbers assigned to respective rectangular regions obtained by dividing a figure of a finger.

FIG. 13 is a diagram illustrating a configuration example of a conventional fingerprint sensor.

[Explanation of symbols]

2 light emitting diode, 3 prism, 4 imaging optical system, 5 CCD, 11 detection unit, 12 optical sensor, 13 pressure sensor, 14 image processing unit, 15
Data collating unit, 15A database, 20F
OP, 20A detection surface, 21 light emitting diode, 2
2 CCD, 26 display unit, 51 detection unit, 52
Image processing unit 53 Fingerprint verification system 54 Registered fingerprint data storage unit 61 Input sensor 62 Binarization processing unit 63 Smoothing processing unit 64 Ridge direction detection unit 65 Ridge direction angle data formation unit 66 Threshold setting section, 67 collation area setting section, 68 registered ridge direction angle data storage section, 69 key input section, 70
Collation processing unit, 71 area setting unit, 72 display device

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] March 25, 1998

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] 0006

[Correction method] Change

[Correction contents]

[0006]

However, fingerprints and palm prints may temporarily change due to cuts, abrasions, burns, etc., in which case fingerprint or palm print verification may not be performed properly. was there.

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0007

[Correction method] Change

[Correction contents]

[0007] The present invention has been made in view of such a situation, and even if a fingerprint or palm print changes due to a cut, abrasion, burn, or the like, it is possible to perform fingerprint verification or palm print verification normally. To make it possible.

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name]

[Correction method] Change

[Correction contents]

Further, without setting the collation area by the key input unit 69, if the collation result of the collation processing unit 70 indicates that only the local part has a low degree of coincidence or that there is only a linear non-coincidence, the collation is OK. It is also possible to do so. Also in this case, it is possible to use a mode switch or the like to indicate that there is a disturbance such as a cut. Further, when there is a thin and wide disturbance such as a scratch, the difference amount between the input image and the registered image is calculated for each area, and the difference amount in each area is equal to or less than a predetermined amount. , Collation OK.

[Procedure amendment 4]

[Document name to be amended] Statement

[Correction target item name] 0091

[Correction method] Change

[Correction contents]

Further, in the third embodiment, the input sensor 61 is constituted by a prism as shown in FIG. 13 , but is constituted by an FOP as shown in FIG. It is also possible.

Claims (8)

[Claims] An input unit for inputting an image of a sample, a matching unit for matching an image input by the input unit with an image of the sample registered in advance, and a matching method of the matching unit, A switching unit that switches according to the state of the sample. 2. The collation apparatus according to claim 1, wherein the switching unit switches a collation method of the collation unit according to a type of the injury of the sample. 3. An input unit for inputting an image of a sample, a conversion unit for converting the image input by the input unit into digital data based on a predetermined reference value, and an instruction for indicating a status of the sample Means, changing means for changing the reference value in accordance with the state of the sample instructed by the instructing means, and detecting means for detecting a characteristic amount of the sample from digital data converted by the converting means. A collation device for collating data representing the characteristic amount of the specimen detected by the detection means with data representing a characteristic amount of the specimen registered in advance. 4. The apparatus according to claim 2, wherein the conversion unit converts the image of the specimen into digital data based on a reference value set for each divided area obtained by dividing the image. Collation device. 5. An input unit for inputting an image of a sample, a specifying unit for specifying a first area where disturbance has occurred in the image input by the input unit, and an input unit for inputting by the input unit. A second area excluding the first area specified by the specifying unit in the image, and a second area of the image of the specimen registered in advance.
A matching unit that matches a region corresponding to the region. 6. An input step of inputting an image of a specimen, a collation step of collating the image input in the input step with an image of the specimen registered in advance, and a collation method of the collation step. A switching step of switching according to the state of the sample. 7. An input step of inputting an image of a sample, a conversion step of converting the image input in the input step into digital data based on a predetermined reference value, and an instruction for indicating a status of the sample A step of changing the reference value in accordance with the state of the sample instructed by the instructing step; and a detecting step of detecting a characteristic amount of the sample from the digital data converted by the converting step. A collation step of collating data representing the characteristic amount of the specimen detected in the detection step with data representing a characteristic amount of the specimen registered in advance. 8. An input step of inputting an image of a sample, and among the images input in the input step,
A specifying step of specifying a first area where disturbance has occurred; a second area of the image input in the input step excluding the first area specified in the specifying step; A collation step of collating an image of the registered specimen with an area corresponding to the second area.