JPS58106660A - Collating device for seal impression - Google Patents

Abstract

PURPOSE:To reduce the collating time, by picking up external characteristics from a registrated seal impression pattern and a seal pattern to be collated, and discriminating the dissidence immediately when the external characteristics are dissident. CONSTITUTION:A seal impression pattern to be collated 2 for a seal to be checked, and a registrated seal impression pattern 5 for a registrated seal are respectively transmitted to characteristic pickup sections 7 and 8. The section 7, 8 are roughly classified into a center detection function and an external characteristic detection function. The detection of center is done by tracing the profile of the external circumference and calculating the center of gravity of coordinates of traced profile. The detection of external characteristics is done by calculating the distance between the pattern external profile and the center point as to each external profile of each pattern and detecting the number of minimum and maximum change points of the distance. Thus, the number of maximum and minimum points detected at the sections 7, 8 is compared at a characteristic comparison section 9, and when the number of detection of the maximum and minimum points is dissident, it is discriminated as dissidence directly at a dissidence discriminating section 6.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a seal imprint collation machine that mechanically determines whether seal imprints match or do not match.

(Background Art) A conventional seal imprint verification device is shown in FIG. FIG. 1 shows a seal imprint pattern 2 to be verified read by a reading unit 1, moved and rotated by a position matching unit 3, and a matching degree with a registered seal imprint pattern 5 is checked by a verification comparison unit 4. The match determination unit 6 makes a determination, and if necessary, the results are fed and banked to the position matching unit 3 to move and rotate in a direction that improves the match, and then the matching comparison unit 4 checks the match again. Investigate
This device repeats these steps and determines the authenticity of the seal impression based on the maximum value of the degree of coincidence. Therefore, with this device, it takes a very large number of times to obtain a judgment result.
``Position matching'' and ``Verification comparison'' must be repeated. By the way, since "verification and comparison" generally involves scanning the entire pattern and examining whether dots on the mesh correspond to each other one by one, it requires a considerable amount of processing time. Therefore, performing "verification comparison" many times leads to the drawback that the time required for verification is very long. By the way, the procedure for checking seals that is currently performed manually is said to be in the following order: ■Shape ■Thickness ■Typeface ■Character (burr) Stroke, and the verification stage of ■■ mainly uses only the external features of the seal imprint pattern. Therefore, it can be said that internal features are ignored.

However, conventional seal imprint matching machines are wasteful because they always match the entire pattern, which is a combination of external features and internal features, even when °C. That is,
If the registered seal imprint pattern is square and the matched seal imprint pattern is round, if the shapes are different, it is not necessary to match all the patterns one by one, and it is safe to conclude that there is a "mismatch." First, it has the disadvantage that all patterns must be matched, which increases the matching time.

(Problems to be solved by the invention) The purpose of the present invention is to eliminate these drawbacks,
External features such as external shape are extracted from the registered seal imprint pattern and the matched seal imprint pattern as the number of thickest values and minimum values of the distance between the center and the contour point, and if the external features do not match, both seal impressions are extracted. This method performs "coincidence determination" without performing "positional matching" or "verification comparison" of patterns, and will be described in detail below.

(Structure and operation of the invention) FIG. 2 shows a first embodiment of the present invention, in which external features such as external shape are extracted from a verified seal imprint pattern 2 obtained by reading a test seal imprint with a reading unit 1. The feature extraction section 8 extracts external features such as the registered seal imprint pattern 5 and the external shape, and the feature comparison section 9 compares the above extraction results to determine the external shape. If the external shape features are clearly different, the operations of the position matching section 3 and the matching comparison section 4 are stopped, and a judgment of "mismatch" is made via the matching determination section 6, and the difference in the external shape features is confirmed. If it is not clear, the position matching unit 3 moves and rotates the seal imprint pattern 2 to be compared, as in the conventional example, the matching comparison unit 4 checks the degree of matching with the registered seal imprint pattern 5, and the match determining unit 6 After the judgment is made and the results are fed back to the position matching section 3 as necessary to perform movement and rotation in a direction that improves the degree of coincidence, the degree of coincidence is checked again in the comparison and comparison section 4. The process is repeated and the authenticity of the seal imprint to be tested is determined based on the maximum value of the degree of coincidence. The feature extraction units 7 and 8 will be explained. Note that since the feature extractor 7 and the feature extractor 8 are the same, only the feature extractor 7 will be described.

The feature extraction unit 7 is roughly divided into a center detection function and an external shape feature detection function. Center detection involves tracing the contour of the outer circumference of the baton.
This is performed using a known method of calculating the center of gravity of the traced contour point coordinates. External shape feature detection is performed by calculating the distance between the center point and the outer contour point of the baton obtained in the above center detection for each of the outer contour points of the baton, and detecting the number of extremely thick and minimum points where the distance changes. .

FIG. 3 is an example of detecting the external shape feature of a stamp stamp.

In the case of (a) in FIG. 3, it is an example of a circular seal impression, and the number of extremely thick and minimum points is 0. In the case of (b), it is an example of an elliptical seal impression, and the thickest and smallest points are P, , P2. P3. The number of the thickest and smallest fangs detected at each point of P4 is 4. Case (c) is an example of a square seal impression, P,, P2. P3. P4. P
,, P6. Extremely thick and minimal points are detected at eight points P, , P,.

In this way, the number of extremely thick and minimum points detected by the feature extraction unit 7 and the feature extraction unit 8 is compared by the feature comparison unit 9,
If the numbers of extremely thick and extremely small points match, the position matching unit 3 and matching comparison unit 4 are operated, and the matching determination unit 6 determines the degree of coincidence. Further, when the detected numbers of maximum and minimum points by the feature extraction units 7 and 8 do not match, the match determination unit 6 directly determines that there is a "mismatch".

Next, extremely thick and extremely small point detection will be explained. Let the contour points sequentially obtained by contour tracking be P, , P2...Pn, and their XY coordinate positions are (X8.Y,), respectively.

(X2. Y2)・(Xn, Y,), and each contour point P1
．． ■) 2..

If the center of gravity of Pn is Po and the coordinates of the center of gravity are (Xo, Yo), the distance from each contour point to the center of gravity is LK-(X
K-Xo)2+(YK-Y(+) possible (K-1,2
...n). Detection of the maximum and minimum points is started from the point PM where 11N, LM, which is the maximum among the distances obtained by the above formula, extends.

First, initialize the distance value - to the PM point as the minimum value, compare the minimum value with the next contour point PM distance value LM+1, and when the minimum value > LM+1, the new minimum value = LM value. do. In this way, the distances of each contour point are compared in sequence and the minimum value is updated. During this maximum lJS value update, when a point appears whose distance value increases by THL from the minimum value, that is, at a point where the distance LR1 of the contour points PR10 satisfies LRI≧minimum value + THL, the minimum value is given point P
RI becomes the minimum point. Next, the distance value LRI at the point PR1 where the minimum point was determined is newly initialized as the maximum value, and the maximum value is sequentially updated in the same manner as the minimum point detection. During this maximum value update, when a point appears whose distance value is THL or more less than the maximum value, that is, the maximum value is given at a point where the distance LR2K of the contour point PR2 LR≦maximum value - THL. Point PR2 is a very thick point. Next, the minimum value is newly determined as LR2 at this PR20 point, and minimum point detection is performed in the same manner as above. The detection of the thickest point and the smallest point are repeated sequentially, and when the detection point reaches "+1",
Moving to point Pl, the pole detection ends at point PM. Here, THL needs to be selected to be larger than the fluctuation value of the distance between the contour point and the center of gravity due to noise of input bangs, etc.

As explained above, in the first embodiment, the feature extracting section finds the center of each contour of the registered seal imprint pattern and the matching seal imprint pattern, detects the external shape features of both seal imprint patterns, and the feature comparing section Since a rough matching degree judgment is performed, there is an advantage that the time required for matching can be significantly shortened.

In the first embodiment, a method for detecting the external shape characteristics of a seal stamp button has been described, but the present invention can be similarly applied to figures written in templates such as flowcharts. That is, as shown in FIG. 4, even in regular polygons, shapes can be distinguished by detecting the number of extremely thick and minimum points.

Figure 4 (a) is a circle with 0 poles, Figure 4 (b) is an equilateral triangle with 6 poles, and Figure 4 (c) is a square with 8 poles.
, Figure 4) has a rhombus with 8 poles, and Figure 4 (a feather has a regular hexagon with 12 poles.Here, the square in Figure 4(C) and the rhombus in Figure 4) have 8 poles. are the same number, 8, but can be distinguished by comparing the distance to the center of the point that takes the thickest value. In other words, in Fig. 4 ←→, P2= P
4= P6= P8, but in Figure 4), 1) 3
=))7gacP, = P5, which can be distinguished from the relationship P3>PI. Furthermore, from the relationship between the position of the thickest point or the smallest point and the center position, it is also possible to detect the angle of inclination of the figures, except for circular ones. As an example, considering the case of FIG. 4 (b), the horizontal position of the point P1 is X7, the vertical position is Y, and the horizontal position of the center point is X. , the vertical position is Y.

Then, the inclination angle θ from the horizontal direction of the apex P1 of the equilateral triangle is: Since a match is determined, it can be used in a seal imprint matching machine that requires a short time for matching.Also, since it is possible to identify the external shape of a figure, it can be used in a flowchart deciphering device, etc.

[Brief explanation of the drawing]

Figure 1 is a block diagram of a conventional seal imprint verification machine.
Fig. 2 is a block diagram of an embodiment of the seal imprint collation machine according to the present invention, Fig. 3) to J are examples of detecting very thick and minimal points of a seal imprint pattern, and Fig. 4 (a) to d) are examples of detecting very thick and minimal points of a seal imprint pattern. This is an example of detecting extremely thick and minimum points of a rectangular figure. 1...Reading section, 2...Seal imprint pattern to be verified, 3
...Position matching section, 4...Verification comparison section, 5...
Registered seal imprint pattern, 6... Match determination section, 7...
Feature extraction unit, 8... Feature extraction unit, 9... Feature comparison unit. Patent applicant Oki Electric Industry Co., Ltd. Patent application agent Megumi Yamamoto −

Claims (1)

[Claims] a reading unit that reads the seal imprint to be verified using an electrical signal; a position matching unit that aligns the position of the seal imprint pattern to be verified read by the reading unit with the registered seal imprint pattern; and a verification comparison unit that performs a verification comparison between the two seal imprint patterns; In a seal imprint matching machine having a matching determination unit that determines whether the two seal imprint patterns match or do not match according to the output verification result and instructs the position matching unit to readjust the position matching, A feature extraction section that extracts the features of the seal imprint pattern and a feature comparison section that compares the external features of both seal imprint patterns are provided. Features of the external shape are extracted based on the detected number of extremely thick values and minimum values of the distance between each point on the outline, and when the detected numbers of both seal imprint patterns do not match, the feature comparison section immediately compares the two imprint patterns to be mismatched. A seal imprint verification machine characterized in that it is configured to determine.