JP3628949B2 - Fingerprint verification device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a fingerprint collation device for collating a fingerprint of a finger.
[0002]
[Prior art]
In the progress of the information society and the environment of modern society, interest in security technology is increasing. For example, in the information society, identity authentication technology for system construction such as electronic cashing is an important key. In addition, research and development is also active in authentication technology for protection against theft and unauthorized use of cards (for example, Yoshima Shimizu et al., A study on IC cards with functions with personal authentication, IEICE Technical Report, Technical report of IEICE, OFS92-32, p25-30 (1992)).
[0003]
There are various authentication methods for preventing unauthorized use such as those using fingerprints or voiceprints. Among them, many techniques have been developed for fingerprint verification technology. In fingerprint collation, a fingerprint sensor that can input a fingerprint image and a collation device that collates a fingerprint image captured by the fingerprint sensor with registered feature information are required.
[0004]
In this fingerprint verification, in order to perform high-speed verification while maintaining practically sufficient reliability, image processing is performed on the fingerprint image, feature points such as Y-shaped branch points and end points are extracted, and predetermined points centered on the acquired vertex There is a technique that uses an area as a window for verification. According to this, it is possible to flexibly cope with a difference in how force is applied at the time of input or distortion of a fingerprint image caused by moving the finger while the finger is attached to the surface of the fingerprint sensor.
However, it is difficult to identify and eliminate fingerprints having similar feature points.
[0005]
However, in the vicinity of the center point of the fingerprint, such as the center of the vortex of the spiral fingerprint, the pattern shape of the ridge line (convex and mountain portion) and valley line (concave and valley portion) is complex, and in many cases each person There is a strong tendency to be different.
Using this tendency, a fingerprint center point called “navel” is detected, a fingerprint image is extracted from a predetermined area centered on the center point, and the registered feature information is collated based on the extracted fingerprint image. Techniques to do this have been proposed. According to this technology, it is possible to improve collation accuracy without slowing the collation speed as compared with fingerprint collation using only feature points.
[0006]
The “navel” of the fingerprint is obtained, for example, as follows. The read fingerprint image is processed into a line image, and as shown in FIG. 3, first, a point S0 having fingerprint pattern information as a line image is set as a start point, and a fingerprint line immediately below the point S0 is selected. The maximum point S1 of the fingerprint line is detected. Next, a fingerprint line immediately below the point S1 different from the fingerprint line having the maximum point S1 is selected, and the maximum point S2 of this fingerprint line is detected. Similarly, when the maximum point of the fingerprint line immediately below is detected, the fingerprint line immediately below the maximum point different from the fingerprint line having the maximum point disappears. A maximum point 301 in a state where there is no fingerprint line immediately below the maximum point is defined as a “navel” of the fingerprint.
[0007]
[Problems to be solved by the invention]
However, the arcuate fingerprint as shown in FIG. 4 has a problem that it is difficult to detect the center (navel) of the fingerprint, and the above technique cannot be applied. In addition, in an actual fingerprint, there are not always many feature points in the vicinity of the “navel”, and there is a problem that the accuracy of collation is not improved even if a small region including the “navel” is targeted for collation.
[0008]
The present invention has been made to solve the above-described problems, and an object of the present invention is to improve the accuracy in fingerprint matching.
[0009]
[Means for Solving the Problems]
The fingerprint collation device of the present invention includes a fingerprint sensor that reads a fingerprint pattern as image data, a fingerprint dictionary in which a part of the fingerprint image data read by the fingerprint sensor is stored as registered fingerprint image data, and a fingerprint sensor A feature point extraction unit that extracts feature points whose pattern changes discontinuously from the image data of the read fingerprint and a determination area having a preset area, and there are the most feature points in the determination area In this way, the area determination unit for determining the position of the determination area in the image data and the data in the determination area at the position determined by the area determination unit from among the image data read by the fingerprint sensor in recording the registered fingerprint image data a dictionary registration unit for storing the registered fingerprint image data to the fingerprint dictionary, the registered fingerprint image data stored in the fingerprint dictionary, the finger in the fingerprint verification It is obtained so as to comprise a comparison unit for comparing and collating the data of the determination area in a position where the area decision unit decides from among the image data read in the sensor.
According to the present invention, a part having many fingerprint fingerprint feature points is compared and verified as a comparison target.
[0010]
In the above invention, the image processing apparatus further includes a fingerprint image processing unit that generates processed image data by binarizing image data read by the fingerprint sensor with a predetermined threshold, and the feature point extracting unit extracts feature points from the processed image data. The dictionary registration unit may extract the registered fingerprint image data from the processed image data in the fingerprint dictionary. The fingerprint image processing unit may generate a processed image by thinning the binarized image data.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a configuration diagram showing a configuration of a fingerprint collation apparatus according to an embodiment of the present invention. The fingerprint collation apparatus firstly includes a fingerprint sensor 101 that reads a fingerprint pattern as image data, and two read image data. And a fingerprint image processing unit 102 for performing a digitization process. In addition, the fingerprint collation apparatus has a feature point extraction unit 103 that extracts a feature point in a fingerprint image binarized by the fingerprint image processing unit 102, and a feature point state extracted by the feature point extraction unit 103. An area determination unit 104 that determines the position of a determination area provided in advance is provided.
[0012]
In addition, the fingerprint collation apparatus stores the fingerprint dictionary 105 that stores and holds the fingerprint image in the determination area determined by the area determination unit 104, and the dictionary that stores the fingerprint image in the determined determination area in the fingerprint dictionary 105 A registration unit 106. In addition, a fingerprint collation unit 107 that compares and collates the fingerprint image stored in the fingerprint dictionary 105 with the fingerprint image in the determination area determined by the area determination unit 104 at the time of collation.
[0013]
The fingerprint image read by the fingerprint sensor 101 and processed by the fingerprint image processing unit 102 is obtained as a binarized image in which the unevenness of the fingerprint is shown in black and white, for example, as shown in FIG. Hereinafter, recording of registered fingerprint image data using the binarized image will be described. First, when the fingerprint image shown in FIG. 2 is obtained from the fingerprint image processing unit 102, the feature point extraction unit 103 extracts the points where the fingerprint lines constituting the fingerprint image are broken and branched as feature points. To do.
[0014]
When the feature point extraction unit 103 extracts feature points, the region determination unit 104 determines the position of the determination region 210 so that there are the most feature points extracted in the determination region 210. The determination area 210 is prepared in advance in the area determination unit 104 and is set to a predetermined area. As a result, for example, as shown in FIG. 2, there are four feature points 203 to 206 in the determination area 210. Finally, the dictionary registration unit 106 registers the fingerprint image in the determination area 210 determined by the area determination unit 104 in the fingerprint dictionary 105 as a registered fingerprint image.
[0015]
The fingerprint image shown in FIG. 2 is registered in the fingerprint collation apparatus as described above, and the fingerprint collation of the user who obtained the fingerprint image shown in FIG. 2 can be performed by the fingerprint collation apparatus. . Conventionally, since the region 230 centered on the “navel” 220 is the determination region, in the case shown in FIG. 2, there are only two feature points 201 and 202, but in comparison, In the determination area 210, four feature points are included, and more feature points are objects of determination in collation.
[0016]
Next, the fingerprint verification of the user will be described. When the user who performs the fingerprint verification causes the fingerprint sensor 101 to read the fingerprint (pattern) of his / her finger to be verified, the fingerprint image processing unit 102 reads the fingerprint. Binarize the fingerprint image. Next, the feature point extraction unit 103 extracts feature points from the binarized fingerprint image. When the feature point extraction unit 103 extracts feature points, the region determination unit 104 determines the position of the determination region so that there are the most feature points extracted in the determination region.
When the region determination unit 104 determines the determination region, the fingerprint collation unit 107 compares the fingerprint image in the determination region with the registered fingerprint image stored in the fingerprint dictionary 105, thereby reading the fingerprint sensor 101. Check the fingerprint image.
[0017]
Although the fingerprint image read by the fingerprint sensor 101 is binarized by the fingerprint image processing unit 102 in the above embodiment, the present invention is not limited to this. The pattern image read by the fingerprint sensor 101 may be thinned in addition to the binarization process. The thinning process is a process of drawing a line passing through the center of the black portion of the binarized image data. This thinning process makes it easier to extract feature points in the feature point extraction unit 103.
[0018]
【The invention's effect】
As described above, according to the present invention, a portion having a large number of fingerprint pattern feature points is compared and verified as a target for comparison and verification, so that an excellent effect of improving accuracy in fingerprint verification can be obtained. .
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a configuration of a fingerprint collation device of the present invention.
FIG. 2 is an explanatory diagram showing image data indicating a fingerprint pattern read by a fingerprint sensor 101;
FIG. 3 is an explanatory diagram showing a fingerprint for explaining a method of extracting a “navel” of the fingerprint.
FIG. 4 is a diagram showing the shape of an arcuate fingerprint.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 101 ... Fingerprint sensor, 102 ... Fingerprint image processing part, 103 ... Feature point extraction part, 104 ... Area | region determination part, 105 ... Fingerprint dictionary, 106 ... Dictionary registration part, 107 ... Fingerprint collation part.

Claims (3)

A fingerprint sensor that reads the fingerprint pattern as image data;
A fingerprint dictionary in which a part of fingerprint image data read by the fingerprint sensor is stored as registered fingerprint image data;
A feature point extraction unit for extracting feature points whose pattern changes discontinuously from the fingerprint image data read by the fingerprint sensor;
A determination region having a predetermined area, and a region determination unit that determines the position of the determination region in the image data so that the feature points are present most in the determination region;
In the recording of the registration fingerprint image data, dictionary for storing data of the determination area in the region determination unit has determined the position from among the image data read in the fingerprint sensor as the registered fingerprint image data to the fingerprint dictionary A registration department;
Comparison for comparing and comparing data in the determination area at a position determined by the area determination unit from image data read by the fingerprint sensor in fingerprint verification and registered fingerprint image data stored in the fingerprint dictionary A fingerprint collation device comprising a collation unit. The fingerprint verification apparatus according to claim 1, wherein
A fingerprint image processing unit that binarizes image data read by the fingerprint sensor with a predetermined threshold value to generate processed image data;
The feature point extraction unit extracts the feature points from the processed image data,
The dictionary registration unit stores the registered fingerprint image data in the processed image data in the fingerprint dictionary. The fingerprint collation apparatus according to claim 1, wherein the fingerprint image processing unit generates the processed image by thinning binarized image data.

Images