JPH0452974A - Fingerprint pattern sorting device - Google Patents

Abstract

PURPOSE:To realize stable and correct pattern sorting by providing a sorting means to sort a fingerprint pattern into plural predetermined fingerprint assortments according to the shape of the ridge line and the tendency of the ridge line direction of a collected fingerprint picture. CONSTITUTION:The fingerprint picture is collected by a picture input part 10, and is photoelectric-converted, and is outputted as two-dimensional picture data by an A/D converting part 11. Then, in a sorting processing part 18, this two-dimensional picture is divided into plural grid-shaped areas, and the ridge line direction of the fingerprint in a corresponding area is detected for every area, and the center position of the finger print is detected from the ridge line direction, and the shape of the ridge line is discriminated from the ridge line direction and the center position, and the tendency of the ridge line is discriminated from the ridge line direction and the center position, and the fingerprint is sorted into plural predetermined fingerprint assortments according to these of the shape of the ridge line and the tendency of the ridge line direction. Thus, the stable and correct fingerprint sorting can be realized.

Description

DETAILED DESCRIPTION OF THE INVENTION BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fingerprint pattern classification device, and more particularly to a fingerprint pattern classification device for automatically classifying the patterns of collected fingerprints.

Prior Art Conventional classification methods using features obtained from ridge directions include matching with a standard pattern created in advance (Goto, Nakamura, Minami, Okuno; "Classification of fingerprint images based on directional distribution patterns"; IEICE Technical Report vol. 1.81. no. 2
10. IE81. -48), or by considering the ridge as a vector field and using the line integral of its rotation and area integral of its divergence (
Noda, Keteru, Kobayashi, Kato; “Fingerprint pattern classification machine”,
IEICE, I]R1, 73-97) A digonal shape formed in a diagonal shape (hereinafter referred to as a delta) is formed by the center (hereinafter referred to as the core) and ridges of the fingerprint, which is the standard for classification. A method has been proposed in which materials are classified based on the presence or absence, number, and location of materials.

"-In the conventional method described above, the presence or absence of deltas plays an important role in the classification process, but deltas are fingerprints."-1
It has the disadvantage that it is difficult to collect depending on its location. In particular, in a fingerprint matching device for criminal investigation that targets latent fingerprints, fingerprints with missing deltas are often processed, and it is usually impossible to use deltas for classification.

Furthermore, the extraction of cores and deltas by standard pattern matching has a problem in that the standard patterns prepared in advance have a large effect on the extraction.

OBJECT OF THE INVENTION Therefore, the present invention has been made to solve the drawbacks of the prior art, and its first objective is to achieve stable and accurate pattern classification using stable delta classification. An object of the present invention is to provide a fingerprint pattern classification device that enables the following.

Structure of the Invention The fingerprint pattern classification device according to the present invention includes means for collecting a fingerprint image and outputting it as two-dimensional image data, and dividing the two-dimensional image into a plurality of regions in a grid pattern (2). means for detecting a ridge direction in that area of the fingerprint; means for detecting a center position of the fingerprint from each ridge direction in the area; and determining a ridge shape from the ridge direction and the center position. a means for determining the ridge direction and a tendency in the ridge direction from the center position; and a fingerprint pattern for each of a plurality of fingerprint classifications predetermined based on the shape of the ridges and the tendency in the ridge direction. It is characterized in that it includes a classification means for classifying.

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention. In this example, a fingerprint image placed on a transparent body or
An image input unit for photoelectric conversion]0, A and D conversion units 11 for converting photoelectric conversion signals into two-dimensional quantized image data, and input of a fingerprint image. An image control section 12 that performs control, an image storage section that stores two-dimensional t'F image data]3, a human control section 1 that inputs information regarding fingerprint images to be classified, and a classification result. and output section 2 () that outputs various messages.
and a classification processing unit 18 that performs classification by inputting image data from the image storage unit 13.

The classification processing unit 18 inputs the image data from the image storage unit 1-3 and the information regarding the fingerprint image to be classified from the human power unit 19 to the output unit 20 via an external interface. --- It is composed of a computer programmer 14, a work memory 17, a memory 16 in which a computer is stored, and a central processing unit 15 that is controlled by the computer.

Regarding the fingerprint input unit 1 [], an optical method using light emitted from a light source against a finger placed on a transparent body is described in Japanese Patent Laid-Open No. 54-49300 and Japanese Patent Laid-Open No. 54-85600. Using boundary conditions [,i':', I TV
There is a device that manually captures a charging conversion image of a fingerprint pattern using an image pickup device such as (tndustrial television), and a device that inputs a fingerprint image of stamped paper 1, which is described in Japanese Patent Application Laid-open No. 55-138174. can be done.

In the present invention, the shape of the ridges (middle row in FIG. 2) is detected for the pattern in the upper row as shown in FIG. This allows for pattern classification. Incidentally, the broken lines in the pattern in the upper part of FIG. 2 indicate the rows where cores are present.

FIG. '3 is a flowchart showing the operation of one embodiment of the present invention. First, in step 310, a fingerprint image is input from the image storage 5 section 1.3 through the external interface 14. Step 31. , I, features are extracted from the ridge direction, and in step 312, the region containing the core is extracted.

FIG. 4 is a diagram showing a method of processing collected fingerprint images in this embodiment, in which the collected two-dimensional image 40 is arranged in a substantially grid pattern,
It is divided into a plurality of regions (hereinafter referred to as zones) 17, and the ridge direction is detected for each zone and processed as zone data 42.

For example, the ridge direction for each zone is 3 as shown in FIG.
Divide 60 degrees into 16 equal parts and label in one of 16 directions (0 to
15) Use the zone data that is the characteristic obtained.

Note that 8 equal parts, 32 equal parts, 64 equal parts, etc. can also be considered.

In FIG. 4, zone 41 is a region containing the core,
Zone 42 is a direction-determined region, zone 43 is an uncertain region where ridges exist but whose direction cannot be determined, and zone 44 represents a non-data region where ridges do not exist.

In the process of extracting features (zone data) from the ridge direction in step 311, it is necessary to extract the zone data shown in FIG. 4, and in step 312, it is necessary to extract the zone 41 including the core. However, these processes are disclosed in Japanese Patent Application Laid-Open No. 138174/1982 and Japanese Patent Application No. 63-099334 and Japanese Patent Application No. 63-099 filed by the applicant.
No. 385 and Japanese Patent Application No. 63-099336, the detailed description is omitted here.

In the next step 313, the shape of the ridge is detected and determined from the zone data obtained from steps 311 and 312 and the zone including the core. The method will be explained using FIG.

Starting from the zone 41 containing the core, within the frame 52, the negative to positive direction of the y-axis (direction axis perpendicular to the finger joints) is the main scanning direction, and the X-axis is centered on the zone 4 containing the core. By sub-scanning in both the positive and negative directions (direction axis parallel to the finger joints), the presence or absence and type of the direction label shown in FIG. 5 is checked only for the direction determination zone 42. This results in
The shape of the ridge on the main scanning line in the y-axis direction is detected.

For example, in the example shown in FIG. 6, the shape of the ridge corresponding to the right flow hoof pattern (No. 2) shown in the middle row of FIG. 2 is detected.

The method for determining the frame 52 in FIG. 6 is as follows. The upper limit of the frame 52 is the line that includes the zone 41 including the core, and from this line as the starting point, the main scanning direction is from the negative to the positive direction of the X axis, and the sub scanning is from the negative to the positive direction of the y axis. The number of labels in each direction of the direction determination zone 42 is calculated, and the horizontal tendency H is calculated using the following formula.

H-[1, QOx a (0) +0.05x (a
(1) + a (7) )+0.05x (a(2)
+a(8) l ] / [1,00x (a(0)
+a(4) ) +0.05x (a(1) +a(
3) +a(5)+a(7) ) +2XO,05X
(a(2) 10a(6) )] where a(n) is the number of direction-determined zones of label n existing on the scan.

Then, the value H calculated above for each scan is compared with a predetermined threshold value Sl, and the row that is the first scan that satisfies the condition H<SL is set as the lower limit of the frame 52.

Note that when the number of sub-scans is an odd number, the lower limit is one line further down in the negative and positive directions of the y-axis.

The left and right limits of the frame 52 are the left and right limits of the two-dimensional image data 40 as shown in FIG. You can exclude it.

The method for determining the tendency of the ridge direction in step 314 will be described based on FIG. 7.8. First, a rectangular frame 61 as shown in FIG. 7 is provided, and only the direction determination zone 42 that exists within this frame is targeted.The number of direction labels in each direction determination zone 42 is calculated, and based on that value, the direction Calculate the trend in the linear direction.

For example, the tendency in the direction from the lower left to the upper right within the frame 61 is calculated using the following formula.

L-[1,00x a (2) +〇,05x (a
(],) + a (3) )+0.05x fa(
0) +a(4) l ] / [1,OOx (a(
2) +a(8) ) +0.50X fa(1) +
a(3) +a(5)+a(7) ) +0.05x
(a(0) +a(4) l X2) By comparing the value calculated in this manner with a predetermined threshold value, the tendency of the ridge direction is determined.

The upper and lower limits of the frame 61 are the same as those of the frame 52, and the left limit is the column containing the zone 41 including the core, and the right limit is the column next to this column in the positive direction of the X axis. be.

Next, two rectangular frames 62 and 63 as shown in FIG. 8 are provided, and calculations similar to those described above are performed for each frame to determine the tendency of the ridges in each parentheses in the J direction. In this example,
- Step; The tendency of the ridge direction corresponding to the J-zu right flow hoof pattern (No. 2) is determined.

The frames 62, 67, and 3 in Figure 8 are determined as described above.
be. The left limit of the frame 02 and the left limit of the frame 63 are determined by dividing the width from the upper limit of the frame 2 in FIG. The 1-limit of the frame 62 is the 1-limit of the frame 52, and the lower limit of the frame 0゛3 is the T; eye of the frame 52.

1-ζ, the main scan from the negative to the positive direction of the y-axis and 1. Then, perform sub-scanning in both the negative and negative directions of the X-axis centering on the zone 711 where 7A is created. ■Determine.

V = [1,00x a (4) →
-0,50X f a (3) + a
(5) )+-0,05x (a(2>+a(f
i) ) ] / [1,,00x ta(0)
→-a(4) l +0.50x fa(1
) +a(3) +a(5)+a(7) ) 4
-2X0.05x (a(2) +a(B) ) ]6
For the name 5L, the value V calculated by the above formula is compared with the predetermined threshold S2, and the column that will be the first eyebrow that satisfies the condition of V<52 is determined as the left limit and right limit of the frame 6263. There is a “C” to do.

In this way, in step 314, the construction method shown in Figure 7 and Figure 8 is used to divide the construction into two stages (the tendency of the ridge toward the
3. The reason for this is as above.

That is, first, the tendency of the ridge direction determined using the frames 62 and 6 is the 4" normal arcuate pattern (No. 5) shown in FIG.
Decide whether it corresponds to 7, and if it is not 1゜,
Does the ridge obtained using frame 61 have the same tendency?
, l to N0.4, and determine the final tendency of the ridge direction (J).

In step 315, "C classification" is performed by comparing the shape of the ridge detected in step 313 and the tendency of the ridge direction determined in step 314 to that shown in FIG.

In step 316, the uncertainty zone 43 is used to determine whether the input fingerprint image is of low quality. This determination method will be described based on FIG.

First, in the classification in step 315, the normal arcuate pattern (No.
, 5) For input fingerprint images classified other than 5), the 6th
A frame 52 as shown in the figure is provided, and step 31 is performed within the frame.
Similarly to the detection of the ridge shape in step 3, the zone 41 including the core is scanned from the starting point (2, in the direction of +1 (from the negative side of the Y axis).

The target is only the uncertain scene 43, and for the human fingerprint image classified as the right loop pattern (NO, 2) in step 3I (), the scene 4 including the core is selected. The direction of the IF of the X coordinate is used as the sub-scan as the center point i, and in step 315, the core is created for the human fingerprint image classified as the left loop pattern (No. 3). ], and set the negative direction of the X coordinate as the sub-scan, and perform step 3 code.
Regarding the human fingerprint image classified as a protrusion arcuate pattern (No, 4) in 5, °C is centered on the zone 41 surrounding Kono', and the positive and negative directions of the ζX coordinate are used as sub-scans, and each y-axis direction is Calculate the proportion of the uncertainty zone 43 that cooperates with main scanning,
. . . Determine whether or not the input fingerprint image is of low quality based on the number of main scans in which the uncertainty zone 43 exists at a certain rate or more (hereinafter referred to as a defective column).

Also, in step 315, a normal arcuate pattern (No, 5) is selected.
For the input fingerprint image classified as 2 as shown in FIG.
In the two rectangular frames 62.6-3, it is determined whether the quality is low or not based on the proportion of the uncertain zone 43 that exists within each frame.

In step 317, it is determined whether the low-quality human fingerprint image determined in step 31.8 is a pair ψ of classification rejection (hereinafter referred to as "reject") or a target of reclassification. In this method, the determination is made based on the number of defective columns used in step 316, but in step 3
In step 1.6, the input fingerprint image determined to be of low quality and the normal arcuate pattern determined to be of low quality are used as the target of the image. For the human fingerprint image determined to be rejected in this step 317, the result of this step 31.7 is outputted by the output unit 20 through the external interface 14 in step 31.8.

In step 319, the human fingerprint image determined to be a target for reclassification in step 316 is corrected. The method is carried out using the established relaxation method, which is a known method.
After the correction, steps 312 and subsequent steps are performed again. Regarding this established relaxation method, please refer to "Ra5enfeld: I
TERTIVE METHODS IN IMAGE
ANALYSIs”, Proc, IEEE Con
f, P, R&1. Pl, 4/1.8 (1977).

In step 320, the shape and position of the ridge detected in step 313, and the tendency of the ridge direction determined in step 3I4 are correlated with those in FIG. 9, so that the input fingerprint image to be classified is It is determined whether there is a second classification candidate.

If it is determined in this step 320 that the input fingerprint image to be classified does not have a second classification candidate, the classification result of step 315 is output from the output unit 20 through the external interface 4 in step 318.

In step 321, a second classification candidate is determined for the input fingerprint image to be classified that is determined to have a second classification candidate. This method consists of the classification result of the input fingerprint image to be classified in step 315 (first row in FIG. 9), the ridge shape and its position detected in step 313 (third row in FIG. 9), and step 314. A second classification candidate is determined by associating the determined ridge direction tendency (fourth row in FIG. 9) with the second row in FIG. Note that the arrow in the third row of FIG. 9 indicates the distance between the shape of the detected ridge and the row containing the core.

If it is determined in step 320 that the input fingerprint image to be classified has a second classification candidate, in step 318 the output unit 20 outputs the classification results of step 31.5 and step 32 through the external interface 14.
The second classification candidates determined in step 1 are output.

As described above, in this embodiment, classification is performed based on the position of the core of zone data.

DETAILED DESCRIPTION OF THE INVENTION As described in detail, according to the present invention, classification is performed by detecting the shape of ridges and determining the tendency of the ridge direction using zone data used in existing automatic fingerprint verification systems. This has the effect of enabling stable fingerprint pattern classification.

Furthermore, the classification effect can also be obtained for classification of latent fingerprints in which the existence of deltas is not guaranteed.

[Brief explanation of drawings]

Figure 1 is a system block diagram of an embodiment of the present invention, Figure 2 is a diagram showing the concept of the pattern classification method used in this embodiment, and Figure 3 is a diagram showing the concept of the pattern classification method used in this embodiment.
FIG. 4 is a flowchart showing the operation of the embodiment of the present invention.
The figure shows an example of the zone data of the collected fingerprint image, FIG. 5 shows the ridge direction label in each zone, FIG. 6 shows the detection of the shape of the ridge, and FIGS. FIG. 8 is a diagram explaining the calculation of the tendency of the ridge direction, and FIG. 9 is a diagram showing the concept of the second classification candidate. Explanation of symbols of main parts 10... Fingerprint image input unit 13... Image storage unit 40... Two-dimensional image data 15... Central processing unit 41. ...
...Zone 17 containing the core...Working memory 4
2... Direction confirmed zone 18... Classification processing section 43... Uncertain zone 44...
Non-data zone

Claims (3)

[Claims] (1) A means for collecting a fingerprint image and outputting it as two-dimensional image data, dividing this two-dimensional image into a plurality of regions in a grid pattern, and determining the ridge direction of the fingerprint in each region for each region. means for detecting the center position of the fingerprint from each ridge direction of the region; means for determining the shape of the ridge from the ridge direction and the center position; The method includes means for determining a tendency in the ridge direction from the center position, and a classification means for classifying the fingerprint pattern into a plurality of fingerprint classifications predetermined based on the shape of the ridges and the tendency in the ridge direction. Characteristic fingerprint pattern classification device. (2) comprising means for determining the quality of the fingerprint image according to the number of regions in which the ridge direction has been detected and determined; and means for correcting the image according to a predetermined method when the quality is determined to be low; 2. The fingerprint pattern classification device according to claim 1, wherein the fingerprint classification is performed again using the corrected image. (3) Fingerprint pattern classification according to claim 1 or 2, further comprising means for classifying up to a second candidate when the degree of similarity is high and the fingerprint classification is similar to a plurality of fingerprint classifications during classification by the classification means. Device.