JPH01213764A - Method for preparing dictionary for collating finger print - Google Patents

Abstract

PURPOSE:To surely remove a beared and a sweat gland without using a mask by judging the beared and the sweat gland with an edge point and a branch point a special micro-picture area as a characteristic point. CONSTITUTION:An image from a fingerprint sensor 13 is taken into a CCD camera 14 and accumulated to a frame memory 15. The image is binarized by a binarizing circuit 16, after a fine line processing is executed by a fine line circuit 18, the picture is stored into a fine line image memory circuit 19. A beared detecting circuit 21 and a sweat gland detecting circuit 22 executes the mask processing of the fine line processed partial image, the edge point and branch point in the image are detected, when the edge point and branch point are mutually in the vicinity and the picture element of a raised line part exists in the middle, it is judged that the beared is obtained and when two branch points are in the vicinity, it is judged that the sweat gland points are in the vicinity, it is judged that the sweat gland is obtained. These beared and sweat gland are removed and the partial image of a truly characteristic binary image is registered at a dictionary output circuit 25.

Description

[Detailed Description of the Invention] [Summary] The present invention relates to a method for creating a dictionary for fingerprint comparison, and an object of the present invention is to provide a highly accurate dictionary creation method that can reliably remove beards and sweat glands without using a mask as in the past. Then, mask processing is performed on the partial image to be registered to detect end points and branch points in the image, and if the end points and branch points are close to each other and there are pixels of a ridge between them, it is determined to be a whisker. However, if the branch points are located close to each other, it is determined that they are sweat glands, and an image is recorded from which feature points corresponding to the whiskers and sweat glands are removed.

[Industrial application field]

The present invention relates to a dictionary creation method for registering characteristic parts of partial images for verification in a fingerprint image verification device.

[Conventional technology]

Currently, fingerprint recognition technology has become an extremely important branch of the field of personal identification, and further improvements in recognition performance are desired.

Conventional fingerprint image matching involves extracting a plurality of characteristic partial images from the entire reference fingerprint image and registering them in an external medium (dictionary) in advance as minutiae partial images for each individual. There is a known method in which pattern matching is performed by scanning these on an input fingerprint image.

Registration in the dictionary is performed by extracting the characteristic parts from the binary image of the fingerprint.

[Problem to be solved by the invention]

When extracting characteristic parts from a binary image of a fingerprint, minute branching parts called sweat glands and whiskers that appear on the image pattern and minute lines branching from the ridges of the fingerprint are used to distinguish between the original features of the fingerprint. is irrelevant. (caused by indeterminate factors such as dust and the pressure of fingerprints) cannot be identified as an individual's identifying characteristics, so it is necessary to remove them.

For this reason, conventional methods have been used to remove sweat glands and beard using a special mask. However, with this method of using a mask, it is practically impossible to remove all sweat glands and beards, and there is also the problem that, for example, true fingerprint ridges that are not beards are removed as beards.
A solution was requested.

The problem to be solved by the present invention is to eliminate the problems of the prior art as described above, and to provide a method for creating a dictionary with high precision that can reliably remove sweat glands and beard without using a mask.

[Means to solve the problem]

In order to achieve the above object, according to the present invention, when only the characteristic parts of a fingerprint image for fingerprint comparison are extracted and recorded as a binarized partial image in a dictionary, the partial image is subjected to mask processing. Detects endpoints and branching points in the image, and if these endpoints and branching points are close to each other, and there is a pixel of a ridge in between, it is determined to be a whisker, and the branching points and branching points are If there are sweat glands, it is determined that they are sweat glands, and an image is recorded from which feature points corresponding to the whiskers and sweat glands are removed.

[For production]

Whiskers and sweat glands are determined using end points and branch points in a specific micro-image area as feature points. Since this judgment is made by focusing on a minute area, the presence of an end point within such an area means that there is a possibility that it is a whisker. However, on the other hand, there is a possibility that such an end point is an end point of a portion branched from another image pattern portion in the very vicinity. Therefore, in the present invention, if a pixel of a ridge exists between an end point and a branch point, it is determined that the pixel is a whisker. If there is no pixel of the ridge part in the middle, it means that the end point is separated from the branch point, and therefore it is not a whisker of the image pattern part of interest.

Similarly, the fact that branch points are located close to each other within a minute region is extremely likely to be a sweat gland with an empty space in the center of the pattern.

Therefore, in such a case, it is determined that it is a sweat gland, and it is removed from the true characteristic part that should be registered along with the whiskers.

〔Example〕

Hereinafter, one embodiment of the present invention will be described with reference to the drawings. In the figure, first, image input is performed in step STI. This image input is performed by converting the image from the fingerprint sensor 13 into a CCD.
This is done by capturing the image into the camera 14 and storing it in the frame memory 15. Subsequently, binarization processing is performed in step ST2. In this binarization process, the binarization (image processing) circuit 16
Binarization is performed by This binarized image is stored in the binarized image storage circuit 17. The binarized image is further processed at step ST.
3, line thinning processing is performed. This thinning is performed in a thinning circuit 18, and the thinning image is stored in a thinning image storage circuit 19. This fingerprint image is subjected to mask processing to obtain feature points such as end points and branch points for each small region (step 5T4).

There are many types of feature point shapes as shown in Figure 3, but they are broadly classified into end points and branch points, and depending on their orientation, end points can be divided into four types, for example (a) to (d). , branch points can be classified into eight types, for example (e) to (1). Therefore, (a) to (d) are E-1 to E-4, (
e) Characteristic codes such as B-1 to B-8 can be assigned to (1), respectively. The above feature extraction is performed by the feature extraction circuit 20.

The above operation is the same as in the case of conventional dictionary creation.

According to the feature of the present invention, a step ST5 of detecting whiskers from the image thinned in step ST3, a step ST6 of detecting sweat glands, and an accompanying step ST7 of removing whiskers and sweat glands are added.

Detection of whiskers in step ST5 is performed in the following manner.

FIG. 8 shows a thinned image of a binarized image of a characteristic part of a fingerprint image. In the figure, B indicates a branch point, e indicates an end point, and X indicates an intermediate point.

In the thinned image (Fig. 8) for which feature extraction has been completed, the vicinity of the branch point (B) (iv ) is 1 (black) in the original binarized image (Figure 4), this is considered a whisker. Conversely, if the midpoint X is 0 (broken) in the original binarized image, it is not determined to be a beard. In the latter case, for example, it may be a part of an adjacent pattern or a whisker of an adjacent pattern, but in any case, it is not a whisker of the pattern to be inspected. In addition, in the case of whiskers in adjacent patterns, when inspecting the pattern, X = 1
Therefore, it can be detected reliably. Whisker detection is performed by a whisker detection circuit 21 (FIG. 1). If the midpoint X=1, the coordinates of the end point e of the branch point B1 are recorded as shown in FIG.

This recording is performed in the beard/sweat gland coordinate storage circuit 23.

Further, in step ST6, as shown in FIG. 7, if there is another branch point B in the vicinity of the branch point B within the minute section (ivXiv angle), this is determined to be a sweat gland. In other words, when two branching points exist in the vicinity, their binarized image becomes as shown in FIG. 6, and it is determined that they are not beards but sweat glands. Detection of sweat glands is carried out by the sweat gland detection circuit 22, and the coordinates thereof are stored in the beard/sweat gland coordinate storage circuit 23 as in the case of beard.
Record it with.

The above-described whisker/sweat gland detection work is performed over the entire screen, and the coordinates of all detected whiskers and sweat glands are memorized.

When the detection of beard and sweat glands is completed in this way, step ST7
The feature points corresponding to the whiskers and sweat glands recorded as shown in FIG. 9 are removed from the original thinned image (FIG. 8) (step 5T7). The thinned image with whiskers and sweat glands removed is shown in FIG. As is clear from a comparison between FIG. 8 and FIG. 10, the whiskers and sweat glands in FIG. 8 (indicated by imaginary lines in FIG. 10) have been removed from FIG.

Finally, based on the thinned image recorded in this way (FIG. 10), the small area selection circuit 24 selects a truly characteristic partial image of the binarized image to be registered (as shown in FIG. 11). Step 5T9) and record this in the dictionary output circuit 25 (FIG. 1) (Step STl0).

By the method described above, a partial image (binarized image) having only true feature points is finally registered as a dictionary.

〔Effect of the invention〕

As described above (according to the present invention, when registering a fingerprint image in a dictionary, it is possible to reliably remove sweat glands and beards that obstruct personal identification without using a conventional mask, thereby creating a highly accurate dictionary. This can further improve fingerprint matching accuracy and identification accuracy.

[Brief explanation of the drawing]

1 and 2 are block diagrams and flowcharts showing an overview of the dictionary creation method according to the present invention, FIG. 3 is an explanatory diagram of an example of encoding a feature small region, and FIGS. 4 and 5 are as shown in FIG. 8. Figures 6 and 7 are diagrams showing the thinned image and the corresponding binarized image obtained by enlarging the minute area of the thinned image from which features were extracted, as shown in Figures 4 and 7. Figure 8 is a diagram similar to Figures 4 and 5, Figure 8 is a diagram showing a thinned image from which features have been extracted, Figure 9 is a diagram showing feature points considered to be whiskers/sweat glands in Figure 8, and Figure 10 is FIG. 11 is a diagram showing a thinned image with whiskers and sweat glands removed, and FIG. 11 is a diagram showing a binarized partial image to be registered as a dictionary. 21... Beard detection circuit, 22... Sweat gland detection circuit,
23...beard/sweat gland coordinate memory circuit.

Claims (1)

[Claims] When extracting only the characteristic parts of a fingerprint image for fingerprint matching and recording it in a dictionary as a binarized partial image, the partial image is masked to detect end points and branching points in the image, and these end points and branching points are detected by masking the partial image. If the points are close to each other and there are ridge pixels between them, it is determined to be a whisker, and if the branching points are close to each other, it is determined to be a sweat gland, and the features corresponding to whiskers and sweat glands are determined. A method for creating a dictionary for fingerprint comparison, characterized by recording an image from which points have been removed.