JPH0433074B2 - - Google Patents

Description

[Detailed Description of the Invention] (Industrial Application Field) This invention relates to a character extraction method, and more specifically, reads characters written on a form, and creates a character pattern string for each character area based on the read characters. This paper relates to a character extraction method for separating and extracting characters.

(Prior art) An optical character recognition device (hereinafter abbreviated as OCR) scans characters written on a form line by line.
The optical signal is converted into an image signal by a photoelectric converter and stored in a line buffer. This line buffer is sequentially read out and the character pattern string is separated into character regions, and recognition is performed using the separated character patterns.However, the character extraction method that extracts one character region from the character pattern string has the performance of OCR. greatly affects.

Next, a conventional character extraction method for separating a single character area from character string pattern data stored in an OCR line buffer will be described.

In OCR, one line is scanned from the top to the bottom of the line buffer where the character string is stored,
By sequentially moving the columns in a direction perpendicular to this scanning, the character pattern in the line buffer is read out.
Then, while scanning one row, black dots (character parts are black dots,
A histogram is created by counting the background area (white dots), and the area of one character is determined by referring to the black dot histogram.

FIG. 10 is a diagram showing a pattern sequence using a conventional black point histogram. In the same figure, 10
0 and 101 are character patterns, which are stored in the line buffer in OCR.
102 is a black point histogram of the character patterns 100 and 101 in the column direction. Also, in the same figure,
Reading starts from the specified position at the left end of the line buffer, and while reading one column, a black point histogram is created for that column, and this histogram is compared with a threshold value α (α: constant), and the column whose histogram is larger than α is The starting point was taken as the end point, and the row smaller than the threshold value α was taken as the ending point, and the area from the starting point to the ending point was cut out as a region of one character.

(Problems to be Solved by the Invention) However, in the conventional method described above, in the case of handwritten characters, the person writing the characters is slanted, or the characters are written outside the character writing frame. There has been a problem in that, due to reasons such as someone hitting a part of a character, adjacent characters overlap and a character pattern of two or more characters is cut out as one character. Furthermore, as can be seen from FIG. 10, since the character patterns 100 and 101 overlap in some parts in the column direction, their black point histogram 102 is formed as one area. Furthermore, if the length from the start point to the end point of the black point histogram is determined and it is determined that there are two or more characters, even if the position corresponding to the predetermined threshold is set as the cutout point, some characters other than the relevant character may be mixed in. There was a problem that some characters were missing.

The present invention is intended to solve these problems, and aims to provide a highly accurate character extraction method with a simple configuration.

(Means for Solving the Problem) In order to solve the above-mentioned problem, the present invention separates a quantized character pattern string obtained by photoelectrically converting a character string written on a form into individual characters. The character extraction method for extracting characters is configured by the following means.

This invention stores a quantized character pattern string in a line buffer memory, scans the line buffer memory for one column corresponding to the column direction of the character string, and creates a black point histogram in the column direction. A means for detecting the width of a black spot histogram in the column direction by sequentially scanning each column, and comparing the width of the black spot histogram in the column direction with a predetermined threshold value to determine how many characters the width of the black spot histogram corresponds to. Based on the detection means and the number of characters included in this sunspot histogram, a region to be subjected to character extraction processing is determined, and a character pattern string within that region is scanned row by row in the row direction to create a blackspot histogram in the row direction. A means for detecting a character circumscribing frame of a character pattern string from a black point histogram in the direction and row direction, a storage means for retaining the character pattern string within the character circumscribing frame, and an edge on the opposite side from the upper and lower sides of the character circumscribing frame a detection means for reading out the content of a character pattern string from a storage means accompanying scanning toward the top side and detecting whether the content is a character part or a background part; The character pattern within the character circumscribing frame is divided into four types: the background part, the character part, and the background part that was not subjected to the scan, when the character part is detected in the scan, and the scanning of the column is aborted. A classification means for classifying a column, a row scanning to detect a change point where the classification changes, store it sequentially, simultaneously hold the state before and after the change point (classification result), and detect the transition of the state as a change in a predetermined classification. The present invention is comprised of a change point detecting means for comparing a combination of transitions and detecting a matching change point, and a means for determining a character cutting position based on the detected change point.

(Operation) According to the character extraction method having the above configuration, the operation is as follows.

The quantized character pattern string is scanned column by column to detect the width of the black dot histogram in the column direction, and this is compared with a predetermined threshold to determine the area for subsequent character extraction processing. Ru. Also,
A black dot histogram in the row direction is created by scanning lines within the area, and a character circumscribing frame is determined from the black dot histogram in the column direction and the row direction. Furthermore, the characters are scanned from the upper and lower sides of the character circumscribing frame to the opposite sides, and are classified into the above-mentioned four types of parts. and,
Scan the lines within the character circumscribing frame again, detect the change points where the classification changes, store the change points sequentially, and compare them with the combination of classification transitions and predetermined classification change transitions to find the matching changes. Detect points. As a result, a character cutting position is determined based on the coordinates of the detected change point.

(Example) Hereinafter, an example of the present invention will be described based on the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention. In the figure, 200 is an image signal from a photoelectric conversion unit (not shown), 201 is a line buffer, and 202 is a sunspot histogram creation circuit 2.
20, a circumscribing frame detection circuit 221 and a character determination circuit 222. 203 is a data switching circuit;
04 is a pattern memory, and 205 and 206 are an x counter in the x direction and a y counter in the y direction, which generate addresses for the pattern memory. 207 is a control circuit. 208 is a pattern area classification circuit;
09 is a circuit for detecting a change point from a white point to a black point.
210 is a pattern area change point detection circuit, 211 is a cutout area detection circuit, and 212 to 214 are registers for determining cutout areas.

The operation of this embodiment will be explained below using the block diagram of FIG.

The character string on the form is converted into a binary image signal 200 by a photoelectric converter, and the line buffer 2
It is stored in 01. The following processing is performed under the control of the control circuit 207. The control circuit 207 reads out the image signal stored in the line buffer 201 column by column from the head position of the line buffer 201, advances sequentially through the columns, and ends when all the character pattern data for one line has been read out. Further, the control circuit 207 starts up the black point histogram creation circuit 220 at the same time as reading the pattern data from the line buffer 201 column by column. The black-spot histogram creation circuit 220 creates a black-spot histogram for one column by counting the number of sunspots being read in one column, and stores it in the histogram memory 230 included in the black-spot histogram creation circuit 220 .
The above process is repeated until the black point histogram for one row and all columns is stored in the histogram memory 230, and the process ends.

After creating a black point histogram for one line, the histogram memory 230 in the black point histogram creation circuit 220 is read from the beginning, and blocks are detected by referring to the black point histogram. The control circuit 207 sequentially reads out the black point histogram from the histogram memory 230 in the black point histogram creation circuit 220, and stores the black point histogram and the threshold value α.
(α: constant, however, in this example α=1
), and if the histogram is large, it is selected as a starting point candidate for a block of characters, the storage address of the black dot histograms is sequentially advanced, the columns whose read black dot histograms are larger than the threshold value α are counted, and β
(β: constant, however, in this example, β=2
) If the rows are continuous, the starting point candidate is taken as the starting point. Further, the row continues to advance, and after the starting point is detected, the row whose black point histogram becomes smaller than the threshold α for the first time is set as the end point, and the area indicated by the length from the starting point to the ending point is set as the block. Next, control circuit 2
07 activates the character determination circuit 222 and sets the length of the detected block to threshold values γ ₁ and γ ₂ (γ ₁ and γ ₂ are constants, however, this value is determined from the average width of the character to be read). In the example, γ ₁ =75, γ ₂ =
125). When the length W of the block is smaller than the threshold γ ₁ , the block is determined to be one character, when γ ₁ ≦W ≦ γ ₂ , it is determined to be 2 characters, and furthermore, when W>γ ₂ , it is determined that the block is 3 characters. It is judged to be more than or equal to characters. Further, after the control circuit 207 determines the block, the circumscribing frame detection circuit 221 is activated for the block to detect the circumscribing frame. Furthermore, when the circumscribing frame of this block is detected, the character pattern within the circumscribing frame is transferred to the pattern memory 204. Here, if W > γ ₂ , that is, if the block is determined to be 3 or more characters, the cutting process is performed from the starting point to γ ₂ to separate the first and second characters,
Starting from the resulting cutting point, γ ₂
The cutting process is performed up to the point ``W'', and then the second character and the third character are separated, and so on. Furthermore, in the case of W>γ ₂ , the pattern memory 20 within the circumscribed frame from the starting point to γ ₂
4, and the rest are transferred again when the cutout points between the starting point and γ ₂ are determined by the following process.
Here, as shown in FIG. 2, which will be described later, the left end of the upper side of the character circumscribing frame is the origin, the lower side position is PB, and the right side position is PR.

Next, processing of characters determined to be two or more characters in the above character determination will be explained based on FIG. 1.

The control circuit 207 includes an x counter 205 and a y counter 2 that provide the address of the pattern memory 204.
06 is set at the left end position of the upper side of the character circumscribing frame, the y counter 206 is incremented, and scanning is performed toward the lower side of the character circumscribing frame. Then, the addresses of the pattern memory 204 are set to (x, y) for the X and Y axes, and the values of the x counter and y counter are used, respectively. The contents of the pattern memory 204 at the position indicated by the address are PM(x,y)
It is expressed as In this example, the white point is defined as PM(x,
y) = 0, the black point is PM (x, y) = 1, the white point detected during scanning from the upper side is PM (x, y) =
2. The white point detected during scanning from the bottom side
PM(x,y)=4. Therefore, the pattern memory 204 in this embodiment has a data width of 3 bits for one mesh. In the pattern area classification circuit 208, an address is set at the left end of the upper side of the character circumscribing frame, and the character pattern is read out from the pattern memory 204. When PM (x, y) = 0, (PM (x, y).OR.2) is set as a new PM (x, y) and sent to the pattern memory 20 via the switching circuit 203.
Write to the corresponding address of 4.

When the change point detection circuit 209 from a white point to a black point detects a black point where PM(x,y)=1, the control circuit 207 aborts the scanning of the column and sets the x counter 2.
05 is incremented by one, and the next column is scanned from the upper side of the character circumscribing frame. Also, when scanning starts from the upper side of the character circumscribing frame and reaches the lower side, the scanning of that column is stopped and the next column is scanned. The above scanning is repeated sequentially, and the process ends after processing the rightmost column of the character circumscribing frame. When the scanning from the upper side is completed, the control circuit 207 sets the x counter and the y counter to the left end of the lower side of the character circumscribing frame, and scans from the lower side to the upper side, similar to the scanning from the upper side. Perform processing. However, PM(x,y)=
When it is 0, PM(x,y). OR.4) to PM(x,
y) in the pattern memory 204. Similar to the scanning from the upper side, the process ends after processing the rightmost column. When the above two types of scanning are completed and the patterns within the character circumscribing frame have been classified, the control circuit 20
In step 7, the x counter 205 and the y counter 206 are set at the upper left end of the character circumscribing frame, and a horizontal scan (line scan) is performed to detect a character cutting area.

Here, an example will be used to specifically illustrate the above column scanning. FIG. 2 is a diagram showing a specific example of vertical scanning in this embodiment. In the figure, 100 and 101 indicate character patterns, 103 indicates a scanning direction from the top side to the bottom side, and 104 indicates a scanning direction from the bottom side to the top side. Further, FIG. 3 is a diagram showing the processing result of the column scanning in FIG. 2. In the same figure,
The white point detected during scanning from the top side to the bottom side (the text portion is the black point, the background portion is the white point) is defined as a C point, and the set of C points is defined as a C area. Furthermore, if a black spot is detected during the scanning, the scanning of the column is stopped at that point and the next column is processed. Here, the set of sunspots is defined as area A. Similar processing is performed when scanning from the lower side to the upper side, the white point detected during this scanning is defined as point D, and the set of D points is defined as area D. By scanning twice, white points other than points C and D, that is, the above 2
The white point that has not been scanned in the previous scan is defined as B point, and the set thereof is defined as B area.

Next, detection of a character cutout area will be explained based on FIG. 1.

First, when the pattern area change point detection circuit 210 is activated by the control circuit 207, it reads out character pattern data (classification results obtained by the column scanning) from the pattern memory 204 and performs row scanning within the circumscribing frame. In addition, the pattern area change point detection circuit 210
processes the character pattern data from the pattern memory 204, but while the character pattern data of the point currently being processed is being processed, it holds the character pattern data of the point immediately before that point, and Compare the currently processed character pattern data with the previous character pattern data. The result of that comparison is
If it is determined that there has been a change, the coordinate position of the previous point is detected and held. In other words, PM(x-1,y)
and PM(x,y), and if they are not equal,
Store the X-axis coordinate x-1 in (xREG I) 214. The cutout area detection circuit 211 holds PM(x, y) in the status register when the pattern area change point detection circuit 210 detects the change point. In the cutout area detection circuit 211, the
It has three status registers (not shown) that hold PM (x, y), and the status registers are configured so that when the change point is detected, the contents of the registers are shifted to adjacent registers. There is. moreover,
When the change point is detected and the shift of the state register is completed, it is detected whether the contents of the three types of state register match the following states stored in the control circuit 207. Set the status register to ST1,
If ST2 and ST3, ST1=4 and ST2=0
and ST3=2 or ST1=2 and ST2=
0 and ST3=4 or ST2=2 and ST3
=4 or ST2=4 and ST3=2. However, ST3 is assumed to be the contents of the current coordinate position. Therefore, if the status register matches the position combination, the cutout area detection circuit 2
The decision signal from 11 is provided to (yREG) 212 and (xRES) 213. The contents of the x counter 205 or y counter 206 stored in each register at that time are transferred from each register to y ₁ ,
Output as x ₂ . Further, _x1 is the contents of the x counter 205 stored in (xREG) 214 when the status register ST3 of the pattern area change point detection circuit 210 is at point C or point D.
X ₁ , X ₂ , and Y ₁ are held in registers included in the control circuit 207. If the status register matches the combination, the horizontal scan of that row is discontinued, the Y counter is incremented, and a new horizontal scan of the next row is performed. The runner-up where all the above horizontal scans are completed within the circumscribed frame is x ₁ (xREG), x ₂
The cutting position is determined based on (xREG) and _y1 (yREG).

The pattern transfer method will be described below using the example pattern shown in FIG. 4 in which the cutout position has been determined. Further, FIG. 4 shows the patterns and cutout positions stored in the pattern memory 204 in the block diagram of FIG. 1. In the coordinates, the horizontal axis is the X axis and the vertical axis is the Y axis, and it is assumed that the pattern memory 204 is located in the fourth quadrant.
XM and YM indicate the size of the pattern memory 204, and in this embodiment, XM=YM=
It was set at 128 meters. PR and PB indicate the circumscribing frames of the patterns stored in the pattern memory 204, and are represented by four straight lines: X=O, X=PR, Y=O, Y=PR. In Figure 4, 3
00,301 is a pattern, straight line Y=y ₁ , X=x ₁ ,
X=x ₂ indicates the cutting position. In the pattern memory in this embodiment, data representing one mesh has the structure shown in FIG. In Figure 5, when (1) is 1, it means that the point was a white point during column scanning from the bottom side to the top side, and when (1) is 0, it means that it was a point other than the white point. . Also, (2) is 1
When (2) is 0, it means that the point was a white point during column scanning from the top side to the bottom side, and when (2) is 0, it means that it was a point other than the white point. Furthermore, when (3) is 1, it means a point that is a black point, and when (3) is 0, it means a point that is a white point. Therefore, the pattern data to be transferred is only the data shown in (3). The mesh on the straight line represented by X=0 from the point Y=0
Y by incrementing the coordinate by one
= Transfer pattern data to point PB. After one column transfer is completed, the X coordinate is incremented. The above transfer is repeated for each column, and when the transfer of the column X=x ₁ is completed, the Y coordinates from the next column to the column X=x ₂ are changed.
From _y1 to PB, pattern data is masked and a fixed value of 0 is transferred. The transfer of the pattern 300 ends when the column of X=x ₂ is transferred. pattern 3
It is also possible to transfer the pattern for 01 using a similar method. Furthermore, for data in which one character is included within the circumscribing frame, the pattern within the circumscribing frame can be transferred using a similar method.

Next, the process flow of this embodiment will be explained in detail based on the flowcharts shown in FIGS. 6, 7, and 8. Here, FIG. 6 shows the overall flow, and FIGS. 7 and 8 show flowcharts of classification of pattern areas and determination of cutout areas by two upper and lower scans, respectively. First, the overall flowchart in FIG. 6 will be explained. In S400, a reading operation is started. In S401, one column of pattern data stored in the line buffer is read out, and the first
In the figure, a sunspot histogram is created by the sunspot histogram creation circuit 220 and stored in the histogram memory 230.
Store in. In S402, it is detected that the creation of all the black point histograms for one row has been completed, and the process of S401 is repeated until all the black point histograms for one row have been created. S40
In step 3, the processed characters are managed and the following process is repeated until all characters in one line have been cut out. In S404, the black point histogram is read from the histogram memory, and the starting point and ending point of the black point histogram are detected and set as a block. Also,
The length of the block is compared with threshold values γ ₁ and γ ₂ and the number of characters the block consists of is stored.
In S405, the circumscribing frame detection circuit 2 of FIG.
21, the circumscribing frame of the block is detected, and the pattern data within the circumscribing frame is stored in the pattern memory 204.
Transfer to. In S406, based on the result of determining the length of the held block, if it is one character, the pattern data in the pattern memory 204 is transferred to the output stage and the process proceeds to the next character. If there are two or more characters, perform the following processing. In S407, column scanning is performed from the upper and lower sides of the circumscribing frame to the opposite sides, the pattern area is classified, and the results are stored in the pattern memory. In S408, horizontal scanning (row scanning) within the circumscribed frame is performed, the classification results are read out from the pattern memory, a cutting area is detected, and the cutting position is determined. In S409, the pattern in the pattern memory is transferred according to the cutout position. When all the patterns in the pattern memory have been transferred, the next character is processed.

Next, S407 and S408 in FIG.
Detailed flowcharts of the processing are shown in FIGS. 7 and 8, and the operations thereof will be explained in order.

FIG. 7 shows the flow of classifying character pattern areas and detecting points of change from white dots to black dots. S5
00, when character pattern data is input, S
501 and S502 are initialization, where the x and y coordinates of the pattern memory are set to the upper left edge of the character circumscribing frame, and the value U/D indicating the scanning direction is set to 2 because scanning is performed from the upper side to the lower side. . In S503, the contents of the pattern memory are checked and PM(x,
y)=1 (black point), the process moves to S507,
When PM (x, y)≠1 (white point), S50
4, the contents of the pattern memory are PM (x, y) = (PM
(x, y). OR.U/D). In S505, the value of the y counter is incremented or decremented depending on the scan direction. S506
Then, one column is managed, and the process returns to S503 and the same process is repeated until the processing of one column is completed. PM
If (x, y) = 1 is detected or one column of scanning is completed, the y counter is set to the scanning start point (on the top or bottom edge) in S507, the x counter is incremented in S508, and x S until the counter matches the right edge of the character circumscribing frame (PR+1)
The process from 503 is repeated. If the x counter matches the right end in the scanning, the scanning direction is set from the bottom side to the top side, and the above process is repeated. At this time, U/D is set to 4 in S511. Furthermore, x,
Initialize the y counter. Similar processing is performed for the scan of U/D=4, and when all is completed, the processing shown in the flowchart of FIG. 8 is performed.

Figure 8 shows the flow of detecting change points in a pattern area and determining a cutout area.
Starting from the left end of the upper side, the extraction area is determined. S
In steps 600 and S601, the y counter and x counter are initialized, respectively. In S602, status registers ST1 to ST1 to hold changes in the area during row scanning are set.
Initialize ST3. ST3 indicates the state of the current position, ST2 indicates the area before the current area during scanning, and ST1 indicates the area before ST2. In S603, the contents PM (x, y) of the pattern memory are compared with ST3, and if they match, the process proceeds to S605; if they do not match, the coordinates are a change point, so in S604, ST2 and ST3 are
The contents of are shifted to ST1 and ST2, respectively. S
In 605, the contents of PM (x, y) are shifted to ST3. In S606, it is determined whether the status register ST3 is at point C or point D. If it is at point C or point D, the contents of the x counter are
). In S608, status register ST
Determine the combination of states 1, ST2, and ST3, and proceed to S6
If it matches the combination shown in 16, S611~
By S613, (xREG) is set to x ₁ and x is set to x ₂ .
Store the contents of the counter in _y1 . In S614 and S615, the y counter is incremented, and the process returns to S601 and the same process is performed until it matches the lower side of the character circumscribing frame. However, if the combination in S608 is the latter two items in S616, then in S611, x ₁ plus -1 from the contents of xREG is stored.

In S608, status registers ST1 to ST3
does not match the combination shown in S616, S
609, S610, the x counter is incremented, and the S
The process returns to 603 and the above process is repeated. y in S615
If the value of the counter matches the lower edge of the character circumscribing frame, the cutting point is determined at the values of x ₁ , x ₂ , and y ₁ . 9th
The figure shows an example of a pattern cut out according to this embodiment, and A-A' indicates the dividing position of the pattern.

As explained above, according to this embodiment, even if the character pattern before and after overlaps with the character pattern, the character pattern can be created without missing the character pattern or with part of the character pattern before and after it being mixed in. can be cut out.

Furthermore, in this example, the case where two characters overlap is shown, but even when three or more characters overlap, two characters from the beginning of the overlapping characters are used.
A similar effect can be obtained by sequentially determining cutout points based on characters.

(Effects of the Invention) As explained above, according to the present invention, the background portion is classified into regions according to the scanning direction by performing column scanning from the upper and lower sides of the circumscribed frame of the character pattern toward the opposite sides, Based on the classification results, the lines within the circumscribed frame are scanned to detect the cutout area and determine the cutout position, so that character cutout can be performed with high precision. Further, since this is realized by scanning the circumscribed frame of the pattern and detecting the change point, it can be implemented with a simple circuit configuration. Furthermore, by using the present invention, it is possible to cut out even when adjacent characters overlap, so it is possible to reduce the interval between character entry frames and increase the number of readable characters per line. Therefore, it can handle many types of forms, has a large degree of freedom in form design, and has high performance OCR.
The effect is that it can be realized.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing a specific example of the column scanning of this embodiment, FIG. 3 is a diagram showing the processing results of the column scanning of FIG. 4 is a diagram showing an example of the pattern in which the cutting position in FIG. 2 has been determined, FIG. 5 is a diagram showing the configuration of the pattern memory in this embodiment, and FIGS. FIG. 9 is a diagram showing an example of a pattern when clipping is performed according to this embodiment, and FIG. 10 is a diagram showing an example of a pattern using a conventional black point histogram. 200...image signal, 201...line buffer,
202... Circumscribing frame creation circuit, 203... Switching circuit, 2
04...Pattern memory, 205...x counter, 2
06...y counter, 207...control circuit, 208...
Pattern area classification circuit, 209...White point→black point change point detection circuit, 210...Pattern area change point detection circuit, 211...Cutout area detection circuit, 212...
yREG, 213...xREG, 214...xREG
.

Claims (1)

[Claims] 1. In a character extraction method in which a quantized character pattern string obtained by photoelectrically converting a character string written on a form is separated and extracted character by character, the character pattern string is stored in a line buffer memory. , create a black point histogram in the column direction by scanning the line buffer memory one column in the column direction of the character string, and perform the scanning sequentially for each column to determine the width (threshold value α) of the black point histogram in the column direction.
(α is a constant) means for detecting the length of a row in which larger black spot histograms continue for a threshold value β (β is a constant) or more columns; and comparing the width of the black spot histogram in the column direction with a predetermined threshold value. means for detecting how many characters the width of the sunspot histogram corresponds to, and determining an area to perform character cutting processing based on the number of characters in the sunspot histogram, and cutting the character pattern string in the area in each row in the row direction. means for creating a black dot histogram in the row direction by scanning each time, and detecting a character circumscribing frame of a character pattern string from the black dot histogram in the column direction and row direction, and holding the character pattern string within the character circumscribing frame; A storage means scans the character circumscribing frame from the upper and lower sides to the opposite sides, reads out the contents of the character pattern string from the storage means along with the scanning, and determines whether the contents are a character part or not. a background portion detected by scanning from the upper side obtained by the detection means, a background portion and a character portion detected by scanning from the lower side, and a character portion detected by the scanning; If so, the scanning of the column is aborted, and therefore the character pattern string within the character circumscribing frame is classified into four types of background portions that have not been subjected to the scanning. Scanning is performed to detect change points at which the classification by the classification means changes, and store them sequentially, simultaneously retaining the states (classifications) before and after the change point, and converting the state transition into a combination of transitions of changes in predetermined classifications. A character cutting method characterized by comprising a changing point detecting means for comparing and detecting the matching changing point, and a means for determining a character cutting position based on the changing point detected by the changing point detecting means.