JPS63229583A - Method and device for fingerprint image collation - Google Patents

Abstract

PURPOSE:To remove a whisker which is an unnecessary noise present in a binarized image of a fingerprint and to obtain an accurate thin-line image by detecting and removing thin lines with line width in the binary signal. CONSTITUTION:The fingerprint image detected by a fingerprint sensor 1 is stored in a frame memory 2 and converted by a binarization circuit 3 into the binarized image, which is stored in a storage device 4. A mask storage device 5 is stored with six kinds of masks of 3X3. A mask position designation circuit 6 sends a designated position to the device 4 to supply the contents of 3X3 picture elements at the designated position in the binarization image to a mask processor 7. The processor 7 matches the six kinds of masks in the device 5 in order and when one of the masks is satisfied, it is decided that its picture element is a thin line, i.e. a point on the whisker. When the thin line is decided, a thin-line removing circuit 8 removes the picture element from the binarized image and puts the processing result back in the device 4.

Description

[Detailed Description of the Invention] [Summary] The present invention is directed to 3.
In the fingerprint image matching method and device that performs fingerprint matching based on a line image, by detecting and removing a thin line with a line width of 1 from the binarized image, This makes it possible to remove so-called "whiskers", which are unnecessary noise, and to obtain accurate thinned images.

[Industrial application field]

The present invention relates to a fingerprint image matching method and device used for fingerprint recognition.

In recent years, as computer systems have become more pervasive in society, people concerned have focused on how to ensure the safety of the systems. Many questions have been raised about the security of ID cards and passwords, which have been used up until now to verify identity when entering computer rooms or using terminals. Fingerprints, on the other hand, are considered the most powerful means of identity verification because they have three major characteristics: ``university for all people'' and ``unchangeable throughout life,'' and much research and development has been conducted into simple personal identification systems using fingerprints. is being carried out.

[Conventional technology]

The conventional fingerprint image matching method involves thinning the binary image of the fingerprint to create a thinned image corresponding to the fingerprint pattern, extracting characteristic parts from this, and comparing this with the registered image. There is something that I have decided to do.

[Problem that the invention seeks to solve]

In the above binarized image, there is unnecessary noise, so-called "
There is a beard. However, the above conventional method
Since the binarized image is simply thinned, the "whiskers" remain as thin lines in the thinned image, making it difficult to perform accurate fingerprint verification.

In view of the above-mentioned problems, the present invention aims to provide a fingerprint image matching method and device that makes it possible to remove "whiskers" from a binarized image. The fingerprint image matching method of the present invention selects a line width of 1 from a binarized image.
The present invention is characterized in that a thin line is detected using a mask, and the detected thin line is removed from the binarized image.

The fingerprint '5rA@ verification device of the present invention records a binary image i,
A binary image storage means for storing α and ζ, a mask storage means for storing a mask for detecting a thin line with a line width of 1, and a position in the binary image for checking with this mask is sequentially determined. A mask position specifying means is used to check the contents of the pixel at the position specified here with the mask, and if the pixel has a line width of 1
A mask processing means for determining whether the line is a thin line with a line width of 1? i! The present invention is characterized by comprising a thin line removing means for removing elements from the binarized image.

[For production]

The "whiskers" present in the binarized image are thin lines with a line width of 1 (line width of one pixel). By using a mask, such a thin line with a line width of 1 can be easily found in a binarized image. Therefore, by removing the thus detected thin line with the line width l from the original binarized image, a new binarized image without "whiskers" is created. Therefore, since line thinning can be performed on a binarized image without "whiskers," the thin line image obtained in this way becomes an accurate fingerprint pattern without unnecessary noise.

〔Example〕

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of the main part of an embodiment of the present invention, and FIG. 2 is a flowchart 1 that schematically shows the processing. Below, the first
The specific circuit operation shown in the figure will be described below.

In FIG. 2, image input is first performed in step 7 STI. This image input is carried out by storing in the frame memory 2 one image of a pilgrimage image that has been extracted by the fingerprint sensor 1 (see FIG. 1). Subsequently, in step ST2, binarization processing is executed. This binarization process is carried out by the binarization circuit 3 to process the fingerprint rd stored in the frame J memory 2.
This is a process of binarizing lii t& based on a predetermined threshold value.

Next, the process proceeds to step ST3, and the process of removing "whiskers", which is the most distinctive feature of this embodiment, is executed. In this process, the binarization 1 symbol tα device 4, the mask storage device 5, the "7 space position 1"
h constant circuit 6, mask processing device 7, and fine line removal circuit 8
Do Yoge ζ. This will be explained in detail below.

First, convert the binarized image obtained by the binarization circuit 3 into a binary 1
Hi image record tl! It is stored in the device 4. On the other hand, the mask storage device 5 stores, for example, information as shown in FIG. 3 (+1) to (f).
Six types of 3×3 masks are stored in advance. As these masks, known masks for line thinning used in line thinning (step 5T4) described later can be used. The four types of masks shown in Figure 3(n) to fd) are for detecting diagonally thin lines with a line width of 1;
The two types of masks, ) and tf), are only for detecting thin lines with a line width of 1 in the vertical and 1 vertical directions. Note that in the Raku mask, at least 1 of each of X and Y (IM is not rOJ).

The mask position 1 determination circuit 6 is a circuit that sequentially determines 1fi positions in the binarized image for comparison with the mask, and sends the designated positions to the binarization 1 symbol tα device 4. Then, from the binarized image storage device 4, the contents of 3×3 pixels at the specified position are given to the mask processing device 7 from among the binarized images stored therein. The mask processing device 7 sequentially compares the six types of masks in the mask storage device 5 with respect to the 3×3 pixels at the specified position, and checks whether these masks are satisfied. Then, if one of the above masks is satisfied, the pixel has a line width of 1
It is determined that the point is on a thin line (i.e., a "whisker").

When the mask processing device 7 determines that the line is a thin line, the t+n line removal circuit 8 removes the pixel determined to be a thin line from the binarized image, and the processing result is stored in the binarized image storage device: n
4, and further send a continuation signal to the mask position 1h to move to the next inspection position, even if it is not determined to be a thin line.
The next specified position is sent to the fixed circuit 6, and the next specified position is given to the binarized image storage device 4.

In this manner, once the detection and removal of "whiskers" by the mask has been completed over the entire screen of the binarized image, the process moves on to the thinning process of step ST4. In this thinning process, the Ill line forming circuit 9 performs line thinning on the binarized image after "beard" removal stored in the binarized image storage device 4,
A thinned image corresponding to a fingerprint pattern is obtained. This thinning can be achieved, for example, by using the same mask as shown in Figure 3, and leaving only pixels that satisfy at least one mask, contrary to the above-mentioned "whiskers" removal process. .

As described above, in this embodiment, the "whiskers" in the binarized image are detected as thin lines with a line width of 1, and by removing them, a new binarized image without the "whiskers" can be obtained. Can be done.

In this way, the thin line image obtained from the binarized image without "whiskers" becomes an accurate fingerprint pattern without unnecessary noise, and therefore the fingerprint matching performed based on such a thin line image is also very accurate. becomes.

Note that the mask and algorithm shown in FIG. 3 are just examples, and any algorithm may be used as long as it can detect a thin line with a line width of 1.

〔Effect of the invention〕

According to the present invention, "whiskers", which are unnecessary noises in a binarized image, can be easily removed as thin lines with a line width of 1. Verification becomes possible.

[Brief explanation of the drawing]

FIG. 1 shows the configuration of the main parts of an embodiment of the present invention.
1 and 2 are diagrams schematically showing the processing in the same embodiment.
FIG. 3 fl)) to (') are diagrams showing examples of masks used in the same embodiment. 3... Binarization circuit 4... Binarization image storage device , 5...'7 disk storage device, 6... mask position (h constant circuit), 7... mask processing device, 8... fine line removal circuit, 9... tm line converting circuit.

Claims (1)

[Claims] 1) Binarize the input fingerprint image to obtain a binary image, thin the binarized image to obtain a thin line image, and match the fingerprint based on the thin line image. A fingerprint image matching method comprising: detecting a thin line with a line width of 1 from the binarized image using a mask, and removing the detected thin line from the binarized image. 2) The fingerprint image matching method according to claim 1, wherein the mask is a thinning mask used for the thinning. 3) Binarization means (3) that binarizes the input fingerprint image
and a thinning means (9) for thinning the binarized image obtained by the binarizing means, so as to perform fingerprint verification based on the thinned image obtained by the thinning means. The fingerprint image matching device includes: a binarized image storage means (4) for storing the binarized image; and a mask storage means (5) for storing a mask for detecting a thin line with a line width of 1. a mask position specifying means (6) for sequentially specifying positions in the binarized image to be compared with the mask; and a mask position specifying means (6) for comparing the contents of pixels at the positions specified by the mask position specifying means with the mask. , mask processing means (7) for determining whether the pixel is a thin line with a line width of 1; and removing pixels determined by the mask processing means to be a thin line with a line width of 1 from the binarized image. Fine line removal means (
8) A fingerprint image matching device comprising: 4) The fingerprint image matching device according to claim 3, wherein the mask is a thinning mask used for thinning by the thinning means.