JPS61160181A - Recognizing device of handwritten character - Google Patents

Abstract

PURPOSE:To enable a character entering person to input the optional number of characters without considering the number of regulated digits or the like by finding out the positional coordinates of an input character in a character input area and outputting the information as a pair with a recognition character code. CONSTITUTION:A recognition processing part 3 controls a series of character recognition processing including the input processing and dictionary matching of a handwritten character and the output processing of the recognized result respectively. An input tablet 2 forms a character input area 10, a touch key area 11 and an error lamp 12 on its tablet surface so that the contact of the pen point of an exclusive input pen 13 can be sensed. A touch key area 11 is selected by the pen point of the input pen 13 to key various commands or the like. A character input area 10 inputs a required handwritten character by bringing into point of the input pen 13 into contact with/moving on the tablet surface. A coordinate detecting circuit consisting of a matrix circuit is built in the inside of the tablet surface so as to correspond to the character input area 10 to detect the X and Y coordinates of the pen point.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field of the Invention The present invention relates to a handwritten character recognition device that receives handwritten characters input to an input tablet and performs dictionary matching processing to recognize human characters.

<Summary of the Invention> The handwritten character recognition device of the present invention has a plurality of character frames in which one or more characters can be handwritten input in the character input area of an input tablet, and the handwritten character recognition device is capable of recognizing handwritten characters in the character input area. When recognizing, the recognized character code,
By outputting the character position coordinates as a pair, it is possible to grasp the character frame in which the character is written, and this allows the character entry person to input any number of characters within the character frame. This makes it easier to write characters and improves work efficiency.

<Background of the Invention> Conventionally, this type of character recognition device has been proposed in which a plurality of character frames are set in a character input area of an input tablet for inputting one character to another.

However, in the case of this type of method, the person writing the characters must always be aware of the top, bottom, left and right sides of the character frame when writing the characters, which significantly increases the mental burden on the person writing the characters and also reduces work efficiency. It is decreasing.

Therefore, the inventor of the present invention developed a handwritten character recognition device that solves the above problem by adopting a method of writing a plurality of characters in one character frame.

According to this device, a character entry person can perform a character entry operation on a character frame without being particularly conscious of the left and right sides of the character frame, and the mental burden of the character entry person can be significantly reduced.
It becomes easier to enter characters, and the efficiency of character input work can be improved.

By the way, when introducing this type of handwritten character recognition device into a processing system that handles exclusively numbers, it is possible to write numbers at any pitch within the character frame, but it is also possible to write numbers within the character frame at a predetermined number of digits. When the number input reaches the specified number of digits, the processing target is shifted to the next character frame.

Therefore, the person writing the characters must always be aware of the number of digits in the character frame when inputting numbers by hand, and when inputting numbers that are less than the specified number of digits, they must manually press the space key or press the digits. 0'' must be input by hand and the digits must be aligned, which places a heavy workload on the person inputting characters and is a factor that impedes improvement in work efficiency.

<Purpose of the Invention> This invention provides an input tablet with a plurality of character frames in which one or more characters can be input by hand in the character input area, and in particular when inputting numbers, the digits of the numbers in the character frame are To provide a new handwritten character recognition device that does not require any awareness of numbers and does not require extra operations such as digit alignment.

<Structure and Effects of the Invention> In order to achieve the above object, the present invention provides an input tablet notebook in which a plurality of character frames in which one or more characters can be handwritten are set in a character input area, and a pen tip for the input tablet. an input pen for handwriting characters by touching the
a recognition processing unit that takes in handwritten input characters from the character input area of the input tablet, executes character recognition, and outputs a recognized character code; In search of
The coordinate information was configured to be output as a pair with the recognized character code.

According to the device of the present invention, since the character frame in which the character is handwritten is determined using the coordinate information, there is no need to specify the number of characters to be written in the character frame, and the person writing the characters is aware of the specified number of digits, etc. You can input any number of characters without having to do so. Moreover, since there is no need to input spaces or numbers to match the digits, the work burden on the person filling in the characters is greatly reduced, and it greatly contributes to improving work efficiency, which is an outstanding achievement of the purpose of the invention. be effective.

<Description of Embodiments> FIG. 2 shows an overall schematic configuration of a system for processing numbers, etc., into which the handwritten character recognition device 1 of the present invention is introduced. The illustrated handwritten character recognition device 1 is composed of an input tablet 2 and a recognition processing section 3, and each of these components is electrically connected to a host-side device 4 via a cord line. The host side device 4 is a CRT (Cathode Ray Tube).
5 and a host computer main body 7 comprising a keyboard 6
A printer 8 and a floppy disk device 9 are connected to the computer.

The recognition processing unit 3 includes a CPU (CentralPr
In addition to RAM (Ra
ndam AccessMemory), ROM (
This recognition processing unit 3 is configured with a computer circuit including a memory such as
, the input processing of handwritten characters, a series of character recognition processing including dictionary matching, and further the output processing of recognition results are controlled.

The input tablet 2 has a character input area 10, a touch key area 11, and an error lamp 12 on the tablet surface.
Each of the areas 10.11 is formed so as to be able to sense the touch of the pen tip of the dedicated input pen 13. The touch key area 11 is arranged with a plurality of function keys that are selectively operated with the tip of the input pen 13 and used to input various commands manually. Character input area 10
is for inputting desired characters by hand by bringing the pen tip of the input pen 13 into contact with the tablet surface and moving the pen tip, and a matrix circuit is provided inside the tablet surface corresponding to the character input area 10. A coordinate detection circuit is incorporated to detect the X and Y coordinates of the pen tip in contact with the character input area 10.

In the character input area 10, a horizontally long rectangular character frame 14 is formed in series over a plurality of stages.

Any number of numbers can be input by hand within these character frames 14. After positioning a form with the same character frame printed on the tablet surface, insert the input pen 13 into the character frame. When numbers are handwritten one after another using the computer, each number is sequentially recognized as a character, and the recognized character code and the position coordinates of that character in the character input area 10 are output as a pair to the host device 4. In this case, the recognition processing unit 3 recognizes the constituent strokes of the handwritten input character every time there is a stroke input, checks the mutual positional relationship of each stroke input, and checks that the input character is composed of multiple strokes. After determining whether it is a number or not, cut out the character according to the stroke recognition result and the number discrimination result,
It is transmitted to the host side device 4.

5(a) to 5(d) show examples of cutting out the character "4" composed of multiple strokes, and the first stroke and the second stroke have a predetermined positional relationship. b
), it is recognized as the character “4” and cut out, but
The cases shown in FIGS. 5(c) and 5(d) in which the first stroke and the second stroke 100 do not have a predetermined positional relationship are treated as input errors. In FIG. 5(b), after the first stroke is recognized as "1", the recognition is changed to "4" at the second stroke. Also, Figure 5 (C)
In this case, the first stroke is treated as an error, but the second stroke is recognized as the character "1" and processed.

The input pen 13 has a structure in which, for example, magnetic lines of force are generated at the pen tip when the pen tip touches the input tablet 2 and a predetermined writing pressure is applied. By detecting each stroke of the input character, from writing to writing, each stroke of the input character is obtained as data.

FIG. 3 shows the contents of the memory in the recognition processing section 3, and a program area I5 in the figure stores a series of programs for character recognition processing.

The work area 16 includes a coordinate counter 17, a coordinate storage area 18, a sub-stroke counter 19, a direction code storage area 20, a stroke code storage area 21, a coordinate counter save area 22, and a stroke code save area 23.
The coordinate counter 17 stores the number of coordinate points that the input pen 13 has passed on the input tablet 2, and the coordinate storage area 18 stores the number of coordinate points that the input pen 13 has passed on the input tablet 2.
The substroke counter 19 stores the number of 8-direction codes (shown in FIG. 4) included in the character stroke, and the direction code storage area 20 stores each substroke (described later) that constitutes the character stroke. The stroke code storage area 21 contains a code specifying the type of stroke input, and the coordinate counter save area 22 and the stroke code save area 23 contain codes for each of the coordinate counter 17 and stroke code storage area 21 at a predetermined stage. The contents will be cents each.

In addition, the stroke recognition dictionary area 24 contains numbers rOJ r2J r3J r6J consisting of one stroke.
Standard collation data for r8J, r9J, etc. are stored in the number recognition dictionary area 25, and standard data for collation for numbers r4J, r5J, r7J, etc. consisting of a plurality of strokes are stored, respectively.

FIG. 1 shows the control operation of the recognition processing section 3. First, in step 31, the recognition processing section 3 is initialized, and the contents of the work area 16 of the memory are cleared. In the next step 32, it is determined whether or not there is a command input from the touch key area 11 of the input tablet 2. When a key requesting tablet input is pressed now, the step 32 ``Is there a command input?''" and "Is this an input start command?" in step 33 are both 'YE.
S'', the process proceeds to step 34, and the contents of the work area 16 of the memory are cleared. If another command input key is pressed, step 33 becomes NO'', and the process proceeds to step 35. The command code is sent to the host device 4.

In this way, when the input pen 13 comes into contact with the character input area 10 of the input tablet 2 and a predetermined writing pressure is applied, the determination of "Is there a command input?" in step 36 is NO, and step 37
The determination of "Is it written?" is "YES", and in the next step 38, the subroutine for stroke input processing shown in FIG. 6 is called.

First, in step 71 of FIG.
It is determined whether the writing point is within the character frame 14, and "YES" is determined.
With this determination, the contents of the coordinate counter 17 are incremented by 1, and the pen position coordinates are detected by the coordinate detection circuit in the input tablet 2 and stored in the coordinate storage area 1st (
Step 72.73). When the pen tip moves and a character stroke is drawn within the same character frame 14, the determination in step 74 is "NO", and the process returns to step 71 to determine "is it within the character frame?""YES", and as before, the coordinate counter 17 is incremented and the pen position coordinates are stored (steps 72 and 73).

This processing operation is repeatedly executed every time the pen tip position passes a coordinate point, and when the pen tip is lifted out of the input tablet 2 by completing one stroke input, the determination in step 74 becomes "YES", and step 75 Proceed to.

In step 75, a point where the coordinates of the stroke input abruptly change (for example, a point where it changes by 90 degrees) is detected as a change point, and the character stroke is divided into a plurality of sub-strokes at each change point.

7(1) and (2) show examples of substroke division for characters r2J r5J. In the figures, black dots 26 are writing points, black dots 27 are writing separation points, white points 28 are change points,
Straight lines 29 each indicate a substroke.

Each of the above-mentioned substrokes is converted into a direction code shown by numbers 1 to 8 in FIG. The direction code of each substroke is stored in the direction code storage area 20 in the following step 77.

FIG. 8 shows an example of direction code conversion for the first stroke of the character "5", and in the figure, ■■■■■■■■ indicates the direction code of each sub-stroke.

In this code conversion, when adjacent direction codes are continuous (in the illustrated example, the portions ``■'' to ``■''), a so-called rounding process is performed to make the direction code ``■■■'' associated with a long stroke representative.

Figure 9 shows the direction code string of the character "5" after rounding, and the first stroke has each direction code of ■■■■,
The second stroke is each configured with a direction code of ■.

Next, the stroke input is compared with the contents of the stroke recognition dictionary area 24 in step 78, and it is checked whether there is standard data having a direction code string that matches the direction code string of the stroke input. Then, when matching data exists, the determination of "Does it match?" in step 79 becomes "YES", and in step 80,
A code specifying the character stroke related to the standard data (hereinafter referred to as "stroke code") is set in the stroke code storage area 21, and the process returns to the main routine. On the other hand, if the standard data related to matching does not exist in the dictionary area 24, the result of step 79 is NO'', and in this case, the process returns to the main routine as an input error.

Note that if the pen tip protrudes outside the character frame 14 during the process of drawing a character stroke, the “Is it within the character frame?
" is NO", an error sound is generated at the step knob 81, and the error lamp 12 on the human-powered tablet console L-2 lights up in step 82. In this case, the pen tip is separated from the input tablet 2. Depending on the writing, the next step 83 becomes "YES", and the process returns to the main routine as an input error.

In this way, if the return is due to a human error from steps 79 and 83, step 39 in FIG. 1 is "YES".
Then, in steps 40 and 41, an error sound is generated and the error lamp 12 is turned on, and further, in step 42, the value of the coordinate counter save area 22 (in this case, the initial value is zero) is set in the coordinate counter 17.

On the other hand, if the input is completed and the process returns to the main routine, the determination in step 39 is "NO", and the process then proceeds to step 43 where the stroke code save area 2
It is checked whether the content of 3 is zero. In this case, since this save area 23 is initially zero, step 43 becomes YES'. Next, in step 44, it is determined whether or not the stroke input related to the verification match is a pair of strokes, that is, whether or not this stroke input is a character stroke of numbers r4J r5Jr7J made up of a plurality of strokes.

When the determination in step 44 is "NO", that is, this stroke input is a number r made up of one stroke.
OJ r2J r3J r6J r8J r9
When it is determined that it is a character stroke of J (excluding "1"), in step 45, the coordinates of the character input area 10 where the character is located (in this embodiment, the coordinates of the writing point) and the character's A pair of stroke codes and stroke codes are sent to the host device 4, and then in steps 46 and 47 the contents of the coordinate counter 17 and both evacuation areas 22 and 23 are cleared, and the process waits for the next stroke input (step 37
) On the other hand, if the stroke input is a character stroke of the number r4j r5J r7J consisting of multiple strokes, the determination in step 44 is "YES" and the process proceeds to step 48, where the character stroke is, for example, the number rlJ. , itself, is checked to see if it constitutes a character. If the determination in step 48 is "YES", the coordinates of the writing point of the stroke and the stroke code are sent as a pair to the host device 4 in step 49, and the value of the coordinate counter 17 is calculated in step 50. is set in one of the save areas 22 and the contents of the stroke code storage area 21 are set in the stroke code save area 23, respectively, and wait for the next stroke input (step 37).

On the other hand, if the character stroke itself does not constitute a character, step 48 is "NO" and step 5
1, the value of the coordinate counter 17 is transferred to one evacuation area 22,
The contents of the stroke code storage area 21 are respectively set in the other save area 23, and the process waits for the next stroke input (step 52). If there is a next stroke input, step 52 becomes ``YES'' and the stroke input process is executed in step 38. On the other hand, if there is no stroke input within a predetermined time, step 52 becomes ``NO''.
, the determination of “Is time over?” in step 53 is “YES”.
S'', and the process moves to step 40 and subsequent error processing.

When inputting the number "4" by hand on plywood, if we specify the point at which the input of the first stroke is completed as shown in Fig. 5(a), the manual force of the first stroke itself constitutes a character. I haven't. Therefore step 4 above
8 is "NO", and the contents of the coordinate counter 17 and the contents of the stroke code storage area 21 for this stroke input are stored in each retreat area 22, 23 (step 51). Then, when the second stroke of the character "4" is input by hand, the answer in step 52 becomes "YES", and in step 38, the stroke input process according to the subroutine shown in FIG. 6 is executed in the same manner as described above. , in this case, return to the main routine, step 39 is "NO", step 43 is "N"
O'' and the process proceeds to step 54. This step 54 determines whether or not the stroke input is a paired stroke. In this case, the determination in step 54 is ``YES'' and the process proceeds to step 55. Proceed to step 55.
Then, the combination of both stroke inputs of the first stroke and the second stroke is checked against the contents of the dictionary area 25 for number recognition, and standard data having that direction code string and a direction code string that is defeated is present. It is checked whether or not. If matching data exists, the determination in step 56 as to "Does it match?" becomes YES, and in step 57 a subroutine (not shown) is called to check the relative positions of the paired strokes.

In this subroutine, it is checked whether the paired strokes are constituent strokes of the characters r4J r5J r7J, and then it is checked whether the mutual positional relationship of each stroke input matches a predetermined condition. It is something.

Figure 10 shows the coordinate positions of writing points a, c, separation points, d, and change point e for the constituent strokes of the character "4". For example, the y-coordinate value of point C y is the average value of the X coordinate values of points e and b ()'a +'jb)
/ Whether it is larger than 2, ■The y-coordinate value y4 of point d is point e
, average value of the X coordinate values of b (y @ ”)'c )
Whether or not the X coordinate value of point e is smaller than /2,
Whether or not the X coordinate value of point d is smaller than the average value of the X coordinate values of points c and d (x c + 14 )/2. Each condition, such as whether or not it is larger, is checked. Then, when both conditions are satisfied, each stroke input in the pair is the character “4”.
'' and returns to the main routine. On the other hand, if either condition is not satisfied, each stroke input in the pair is determined not to be a constituent stroke of the character "4" and returns to the main routine.

Unique conditions are also set for the characters r5J and r7J, and the same condition check as above is performed.

In this way, when it is determined that the stroke is one of the constituent strokes of the characters r4J r5J r'H, the determination of "Does the mutual positional relationship condition match?" in step 57 becomes "YES", and the character input where that character is located is determined in step 58. The coordinates of the area 10 (in this embodiment, the coordinates of the writing point of the first stroke) and the character stroke code are sent as a pair to the host device 4, and then the coordinate counter 17 and each evacuation area 22, 23 are The contents are cleared (steps 46 and 47).
), and waits for the next character input (step 37).

On the other hand, if the determination in step 56 or step 57 is NO, the process moves to step 59, where it is checked whether the second stroke input itself constitutes a character. If @YES', the process returns to step 45, and the coordinates of the writing point of the first stroke and its stroke code are transmitted to the host side.

If the determination at step 59 is NO''', the process returns to step 40 and the stroke is treated as an error.

Note that the mutual positional relationship of stroke inputs is not limited to the condition check in the above embodiment, but can also be checked by introducing other conditions.

FIG. 11 shows the contents of the memory in the host device 4. As shown in FIG. In the figure, a program area 60 stores a program for receiving and processing commands and character recognition data sent from the handwritten character recognition bag w1, as well as a series of programs related to system processing, and a form format information writing area. 61 stores the format of the form placed on the input tablet 2 (the number of character frames, the position coordinates of the character frames, etc.).

The work area 62 has a received character number storage area 63, a received character storage area 64, a character frame number storage area 65, a received coordinate value save area 66, etc. The received character number storage area 63 stores the number of characters received for one character frame. However, the received character storage area 64 contains the recognized character code and its coordinate information sent from the handwritten character recognition device 1, and the character frame number storage area 65 contains the number of the character frame surrounding the received character, and the received coordinates. Coordinate information that has already been set in the received character storage area 64 at a predetermined stage is set in the value save area 6G.

Furthermore, the host device 4 has an area for storing received information for each character frame, and each area includes a maximum number of digits storage area 67, a character counter 68, and a character storage area 6.
9 is provided. In the maximum number of digits storage area 67, the maximum number of digits of numbers to be stored is set. Character counter 68 counts the number of characters to be stored. In the character storage area 69, received characters are stored one after another at addresses corresponding to the value of the character counter 68.

FIG. 12 shows the control operation of the host side device 4. First, initialization is performed in step 91, and the contents of the work area 62 of the memory are cleared.

Next, the contents of the form format information storage area 61 are read out and the form format is displayed on the screen of the CRT 5, after which reception of information from the terminal (including the handwritten character recognition device 1) is permitted (step 92.93).

At this stage, the handwritten character recognition device 1 transmits the recognized character code and the coordinate information of the character as a pair, and if this is received by the host device 4, step 94 "Received?"
” is determined as “YES”, and the determination in step 95 as to “Is the received content a command?” is “NO”, and the received information is stored in the received character storage area 64 and the number of received characters is “1”.
” is entered in the number of received characters storage area 63, and
Based on the coordinate information received in step 96, a search is made to determine which character frame the received information falls under.

Note that when a command is received in the reception acceptance state, step 95 becomes YES'', reception from the terminal is prohibited in step 97, and then the process corresponding to the command is executed in step 98.

Next, in step 99, it is checked whether the received character is the first character, and in this case, since the determination is YES', the process advances to step 100. In step 100,
The number of the character frame related to the search is set in the character frame number storage area 65, and the contents of the character counter 68 and character storage area 69 of the corresponding character frame are cleared in step 101, and the contents of the character counter 68 and character storage area 69 of the corresponding character frame are cleared in step 102. The character display within the displayed corresponding character frame is also cleared.After that, the coordinate information among the received information is stored in the received coordinate value save area 6.
6 (step 103), and then the first received character is displayed at the head position within the corresponding character frame displayed on the CRT 5 (step 104). Further step 10
In step 5, 1 is added to the contents of the character counter 68, and in the next step 106, the code of the received character is stored in the character storage area 69 using the value of the character counter 68 as a relative address, and then the process returns to step 94. Waits to receive the next information.

Next, when information about the second character included in the same character frame is received, the determination in step 99 "Is it the first character?" is "NO" and the determination in step 107 "Is it the same character frame?"
The determination of j is "YES" and the process advances to step 108.

In step 108, it is determined whether the content of the character counter 68 is smaller than the content of the maximum number of digits storage area 67. If the determination in step 108 is 1YES'', the next step 109 is to determine whether or not the content of the character counter 68 is smaller than the content of the maximum number of digits storage area 67. It is checked whether or not the coordinate information of the received character matches the stored contents of the received coordinate value save area 66. That is, in step 10
Step 9 checks whether the first received character corresponds to the first stroke of the first received character, and if the determination is "NO" (both characters are separate characters), step 110 is performed. The coordinate information of the second received character is set in the save area 66 in step 111, and the received character is displayed on the CRT 5 in step 111.
Then, the character counter 68 is added, followed by step 106.
The respective characters are stored in the character storage area 69.

On the other hand, if the determination in step 109 is "ygs" (both characters are one), then in step 112 the evacuation area 66
Clear the previous coordinate information stored in , and proceed to step 11.
In step 3, the immediately preceding character (in this case, the first character) currently being displayed on the CRT 5 is erased, and in its place, in step 114, the character just received is newly displayed on the CRT 5, and then the process moves to step 106, where the character is memorized. The memory contents of area 69 are rewritten.

The above series of processing is repeatedly executed each time a character is received, and as a result, the received characters are displayed one after another on the CRT 5, the character counter 68 is incremented, and the received information is stored in the character storage area 69. is stored (steps 111, 105, and 106), and when the count value of the character counter 68 reaches the maximum number of digits, the determination in step 108 becomes "NO", and in this case, the count value of the character counter 68 becomes Step 115 to match the maximum number of digits
is “YES”) and the coordinate information of the received character and the evacuation area 6
Only when the coordinate information of the immediately preceding character stored in step 6 matches ("YES" in step 116), step 112 is executed.
The display returns to display the received characters. On the other hand, if the determination in either step 115 or 116 is "NO", the input is over, and the process proceeds to step 117, where the CRT5
An error message is displayed on the screen and an error sound is generated, and in the next step 118, the character counter 68 is incremented by one.

When the reception process for one character frame is thus completed and the character information of the next character frame is received, the corresponding character frame is searched from the coordinate information (step 96). In this case, the number of the searched character frame and the character frame number storage area 6
Since it does not match with the character frame number that has already been marked 5, the determination of "Are they the same character frame?" in step 107 is "NO", and after the new character frame number is cented in step 100, steps 101 and subsequent steps are performed. The following processing is executed.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing the control operation of the handwritten character recognition device according to the present invention, FIG. 2 is an external view showing the overall configuration of the handwritten character recognition device, FIG. 3 is a diagram showing the memory map of the handwritten character recognition device, and FIG. Figure 4 is a diagram for explaining an 8-direction code, Figure 5 is a diagram showing an example of cutting out a character composed of multiple strokes, Figure 6 is a flowchart showing a subroutine for stroke input processing, and Figure 7 (1). (
21 is a diagram showing an example of substroke division for characters r2J r5J, FIG. 8 is a diagram showing an example of direction conversion of the first stroke of character "5", and FIG. 9 is a diagram showing the character "5" after rounding processing. Figure 10 shows the direction code string for the character "4".
FIG. 11 is a diagram showing the memory map of the host-side device, and FIG. 12 is a flowchart showing the control operation of the host-side device. 1...Handwritten character recognition device 2...Input tablet 3...Recognition processing unit 10...Character input area 13...Manual pen 14・・・・・・Character frame

Claims (3)

[Claims] (1) An input tablet with multiple character frames in which one or more characters can be handwritten input in the character input area, an input pen for handwriting characters by touching the pen tip to the input tablet, and an input tablet. and a recognition processing unit that takes in handwritten input characters from the character input area of the tablet, executes character recognition, and outputs a recognized character code, and the recognition processing unit calculates the position coordinates of the input character in the character input area. A handwritten character recognition device configured to output the coordinate information as a pair with the recognized character code. (2) The recognition processing unit includes a number input recognition unit,
means for recognizing the constituent strokes of a handwritten number for each input stroke; means for checking the mutual positional relationship of each recognized stroke input to determine a number consisting of a plurality of strokes; and the stroke recognition result and number discrimination. The handwritten character recognition device according to claim 1, further comprising means for cutting out and outputting handwritten input characters according to the result. (3) The handwritten character recognition device according to claim 1 or 2, wherein the recognition processing section is constituted by a computer including a CPU and a memory.