JPS59782A - Control system of seal impression pattern - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION (1) Technical Field of the Invention The present invention relates to a seal pattern control method, and in particular, to a method for controlling a seal pattern so as to increase the degree of matching when comparing seal imprints. This is what I did.

(2) Prior art and problems For example, when a customer requests a refund of a deposit at a bank, etc.
Alternatively, when requesting a refund by check, etc., the seal stamp affixed to the refund request form or check must match the seal impression registered with the account number written on the refund request form or check. Therefore, in the past, bank employees had to visually compare these seal impressions. Therefore, verification takes time and there are also problems in terms of accuracy.

In order to improve such problems, research is being conducted on using data processing devices to automatically verify seals. That is, as shown in FIG. 1, the registered seal impression pattern is stored in the registered stamp file 5 at the center, and the refund i
Image of stamp pattern 1 stamped on bills, etc.
An optical image input device 2 such as a reader reads it, converts it into a binary value, and sets it in an image memory 3.
The seal imprint pattern set in 1 and the registered seal imprint pattern transmitted from the registered seal file 5 of the center are held in the registered seal data buffer 6, and these are compared by the comparison determination section 4. Then, the degree of coincidence is determined and transmitted to the judgment result output section 7, and the judgment result is outputted by a display section 8 such as an ORT or a printer 9 or the like.

However, in this type of verification, the registered seal imprint and the to-be-verified seal imprint are compared bit-by-bit, but the stamping pressure and vermilion coverage of the imprint pattern are actually not constant, and there may be smearing, blurring, etc. They may exist or have different densities, and these problems degrade the degree of matching. In other words, in the conventional method, when converting into binary information, verification is performed without matching the levels of the registered seal imprint and the verified seal imprint, which was a major cause of a decrease in the degree of verification. .

(3) Purpose of the Invention The purpose of the present invention is to provide a seal pattern control system which improves the above-mentioned problems and prevents the degree of matching from decreasing even if the seal impression is blurred, faded, density difference, etc.

(4) Structure of the Invention In order to achieve this object, the seal pattern control method of the present invention includes an image reader that reads a seal impression and outputs a multilevel density signal, an input holding means that holds the output of the image reader, It is equipped with a histogram creation means for counting the number of pixels for each level, a multi-value level change setting means, and a slice level variable means, and the multi-value level change setting means determines how the multi-value level of the image reader is output by applying a subtraction value to the multi-value level change setting means. The present invention is characterized in that the slice level is automatically changed when converting to a signal.

(5) Examples of the invention An embodiment of the present invention will be described with reference to FIG.

In FIG. 2, 10 is an image reader, 11 is an input memory, 12 is a 4-value setting section, 13 is a histogram creation section, 14 is a verification mark memory, 15 is a data compression section, and 16 is a 2
17 is an "l" level counting section, 18 is an OR circuit, 19 is a transmitting section, 20 is a receiving section, 21 is a data expansion section, 22 is a registration mark memory, 23 is a "1" level storage section, 2
4 is a "1" level comparison section, and 25 is a verification determination section.

The image reader 10 reads out the seal pattern and converts it into a 16-value output signal, and the output data is held in the input memory 11 and transmitted to the histogram creation section 13. And histogram creation part 1
3, the number of data is counted for each output level of 16 values, and a histogram formed of data values for each level of 16 values is created for the seal impression pattern. Then, the 4-value slice level determined by the state of the histogram is output to the 4-value setting section 12, whereby the 16-value stamp pattern transmitted from the input memory 11 is
It is converted into a value and set in the verification mark memory 14.

The data compression unit 15 receives the 4 data transmitted from the verification mark memory 14.
The digitized signal is compressed and encoded to compress the data, and is sent to the center from the transmitter 19 via the OR circuit 18.

At this time, the account number and registered name are also transmitted, but at the same time, the 4-valued signal sent from the verification mark memory 14 is binarized into rlJ and rOJ by the binarization setting section 16 and "1" level counting section 17 The number of bits at the "1" level is counted, and this counted value of "1" is also sent to the center via the OR circuit 18. In this manner, at the time of registration, the four-valued signal, the number of "1" levels, etc. are registered at the center along with the account number, registered name, and so on.

When the seal imprint actually stamped on the refund bill is compared 6, the image reader 10 reads the stamp imprint pattern to be compared, as in the case of registration, and the four-valued seal imprint pattern is stored in the verification stamp memory 14 as described above. At the same time, this is binarized into rlJ and rOJ by the binarization setting section 16, and the "1" level is counted by the "1" level counting section 17,
This is transmitted to the 1-level comparison unit 24. Separately, the account number of the creator of the refund request form is sent to the center, and the center sends the already registered seal imprint pattern signal (compression encoded 4-valued bell signal). ) and the level number “1” are received by the receiving unit 20, and the compressed stamp pattern signal is restored to the original signal by the data decompressing unit 21 and stored in the registered stamp memory 2.
Store in 2.

In addition, the number of "1" levels sent from the center is stored in the "1" level storage section 23, and this is stored in the "1" level comparison section 23.
4. As a result, the "1" level comparing section 24 compares the "l" level number sent from the center and the "1" level number created from the verification marks, and if they do not match, the histogram creating section 13 "1" Level counting section 1
A mismatch signal including whether the number of "1"s set to 7 is large or small is sent, and the quaternization setting section 12 is made to change the slice level. At this time, rlJ level counting section 1
71C When the number of "1"s set is large, the stamp to be verified is stamped darker, and the slice level is changed accordingly.

Then, the number of "1" levels obtained from the binarization setting section 10 is again compared with the number set in the "1" level storage section 23. In this way, this slice level is increased by Xt until both values match within a certain range. When these are approximately equal to 1, the "1" level comparison section 24 sends a matching signal to the matching judgment section 25, and the matching judgment section 25 registers it as the stamp pattern that was set in the matching stamp memory 24 at this time. It compares the stamp pattern with the stamp pattern set in the stamp memory 22 and outputs the result to a display device, printer, etc.

In this way, even when checking is performed in a state where the seal to be compared and the registered seal have different shading in the seal impression padn, accurate comparison can be performed.

Next, a slicing method suitable for seal verification as described above will be explained.

When comparing seal impressions and determining the slicing level of the seal imprint using a histogram, the ability to slice at a fixed level as shown in FIG. 3 has been studied.

In FIG. 3, 16 input from the image reader 30
The input data of the value is set in the input memory 31, and the number of pixels of each level is counted by a counter in the histogram creation section 32 for each level from 0 to F. Then, the fixed slice unit 33 outputs slice values at a preset slice level, and thereby the 16-value input data in the input memory 31 is converted into 4 values by the 4-value setting unit 34, and transmitted to the display memory 35, for example. This will be displayed on the display 36. However, in such a slicing method for creating verification data, the slice level is the same when the seal impression is dark and when it is light, resulting in 4-valued data in which the negative effects of the shading of the seal imprint remain. In other words, when converting into four values (or binarization) at the same slice level, both the dark imprint shown by the A curve in Figure 6 and the light imprint shown by the 8 curves, the dark imprint will be dark and the light imprint will be light. converted, the degree of matching decreases, and it is not practical for operation. Therefore, dark eyes are pale;
It is desirable to automatically set a slice level that darkens light slices.

For this purpose, in the present invention, as shown in FIG. 4, the average slice control section 37 is used to automatically select the slice level according to the shade of the seal impression. This average slice control 37 is selected so that the number of pixels existing between four-value slice levels is approximately 1, when the sum of pixels of the density I-F of the seal impression is M, As shown in FIG. 5, data register 40. Arithmetic unit 41, average value register 42, comparison unit 43, addition unit 44, AND circuit 4
5, an output section 46, etc., and performs the following operations under the control of a control section (not shown).

(1) First, the number of pixels of each density level calculated by the histogram creation section 32 is set in the data register 40, and the calculation section 41 calculates the sum M of the number of pixels of density level - F, excluding level 0. The reason why level O is excluded here is because level O is a white background area and has nothing to do with the shading of the seal impression.

■Next, the calculation unit 41 calculates the average pixel distribution number within each four-value level, which is the total sum M, and stores it in the average value register 42.
Set to . At this time, round down any fractions.

(2) Find the sum of the number of pixels from level 1 that is close to this average value M. That is, level 1. from the f-data register 40.
Sequentially adding the number of pixels of 2... in the adding section 44,
The comparison unit 43 compares the addition result with the Y average value Mh* and controls whether or not to perform the addition.

Then, a town where BH,6<M- is detected. Here L
i indicates the density distribution of t=14 level i, that is, the number of pixels. When this nl is detected, the comparison section 43 outputs "1", turns on the AND circuit 45, and sets the density level n1 in the output section 46. Then n, + l = ? Detect L2 and n3 such that εLZ≦−. and n3+1
Let z−n, 4 = n.

Detect n, M (2), n such that DL$≦-. And town +14
44 = n.

② When converting into four values based on the results of ② to ② above, proceed as follows.

"00" → 0~n.

rolJ -+ n2-n3 rlOJ + n4~n.

"11" → n6 to F When binarizing, "01" and "10" are used as threshold values. In this way, when 4-value data is to be converted using the average slice method, the 4-value conversion setting section 34 performs this process. Further, when converting into a binary value, the 4-value output of the 4-value setting section 34 may be converted into a binary value as described above.

In this way, as shown in Fig. 7, the number of pixels (area in the figure) in each slice level can be made almost equal, and as a result, a dark impression (d is somewhat light, and a light impression becomes somewhat dark and is converted into 4-valued image). /Can be binarized.

Therefore, by incorporating the function of the average slice method as shown in FIG. 4 into the histogram creation section shown in FIG. 2 and using it at the time of registration, it is possible to further increase the degree of matching.

(6) Effects of the Invention According to the present invention, the slice level can be automatically determined according to the shading of the seal impression, so it is possible to register in a state that is easy to check at the time of registration, and it is possible to efficiently check the level at the time of registration. A good match can be made.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram of a conventional seal matching method, FIG. 2 is an explanatory diagram of a seal matching method using the seal pattern control method of the present invention, and FIG.
The figure shows a conventional seal pattern control method, FIG. 4 shows an embodiment of the seal pattern control method of the present invention, FIG. 5 shows a detailed diagram of the average slice control section in FIG. 4, and FIGS. 6 and 7 show the operation. It is an explanatory diagram. In the figure, 1 is an imprint pattern, 2 is an optical image input device, and 3
is an image memory, 4 is a verification judgment unit, 5 is a registration mark file,
6 is a registered mark data buffer, 7 is a judgment result output section, 8
10 is an image reader, 11 is an input memory, 12 is a 4-value setting section, 131d is a histogram creation section, 14 is a verification mark memory, 15rI'i
Data compression section, 16 is binarization setting section, 11-1 rlJ
Level counter, 18 is an OR circuit, 19 is a transmitter, 20 is a receiver, 21 is a data expansion unit, 22 is a registration mark memory, 2
3 (d"1" indicates a level storage section, 241drlJ level comparison section, 25 indicates a collation determination section, and 37 indicates an average slice control section. Patent applicant Fujitsu Ltd. Representative patent attorney Akira Yamatani Eiji 1 Figure 2, Figure q.

Claims (3)

[Claims] (1) An image reader that reads a seal impression and outputs a multilevel density signal, an input holding device that holds the output of the image reader, a histogram creation device that counts the number of pixels for each level, and a multilevel level changer. The apparatus comprises a setting means and a slice level changing means, and is configured to automatically change the slice level when the multi-value level output of the image reader is converted into a predetermined level value signal by the multi-value level change setting means. A stamp control system featuring a turn control system. (2) The stamp pattern control method according to claim (1), characterized in that the slice level variable means includes average slice control means. (3) The stamp pattern control system according to claim (1), characterized in that the slice level varying means includes a "1" level number comparing means.