JPS61163476A - Hand-written character recognition device - Google Patents

Abstract

PURPOSE:To remove distrust and to speed-up character input operation by obtaining the character end point and variation point from each stroke in hand- written input characters, thereby forming compressed images of input characters to inform of the compressed processing step of characters to the person entering the character. CONSTITUTION:In case hand-writing characters by contacting the pen point against the input tablet having character input area, the character recognition control processor introducing hand-written input characters is constituted of CPU, and memories of RAM, ROM, etc., and by detecting the X-coordinate and Y-coordinate of the pen point contacting the input area, displays the character on the CRT5, the input image of the character on 5B and the compressed image on 5C. The compressed image 5C, after obtaining the variation point of the input character by the uniform division method or variation degree detection method from the stroke of the input character, is formed by joining the character end point and variation point linearly and displays the same on the CRT5 screen together with the input image 5B of the input character.

Description

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

<Summary of the Invention> In the handwritten character recognition device of the present invention, a compressed image generated for character feature extraction is displayed on a display unit for an input character handwritten on an input tablet,
This is intended to help the person filling in the characters understand the character recognition process inside the device, so that the person filling in the characters can input characters that match the characteristics of the device.

<Foreground of the Invention> In a conventional handwritten character recognition device, when a character is handwritten input into the character input area of an input tablet, a control processing unit takes in the input character and executes character recognition, and also inputs the character. It is configured to display an image representing the font as it is (hereinafter referred to as an "input image") on a display unit such as a CRT.

By the way, the control processing unit described above determines character end points and change points from each stroke of the input handwritten character, and then creates a compressed image (hereinafter referred to as After generating a compressed image (referred to as a "compressed image"), the geometric features of the input characters are extracted from this compressed image and matched against standard patterns (dictionary matching).
is moving to.

Figure 13 shows the input image (left side figure) and its compressed image (right side figure) for the number "2". In the figure, white circles 51 and 52 indicate the end points of the character, and black circles 53 indicate the change points, respectively. However, in the case of the compressed image shown in the example, the number "2"
'' rather than the alphabet rZJ, and therefore there is a very high possibility that the control processing unit will misread this as the alphabet rZJ.

Thus, in the case of the example shown in FIG. 13, the input image of the input character "2" and the misread recognized character
ZJ will be displayed, but if the person writing the text is not familiar with the internal operation of the device (especially the compression processing operation mentioned above), he or she may be able to check the performance of the device by looking at the discrepancy in the display contents. etc. (resulting).In addition, since the cause of the misreading is unclear to the person writing the letters, he or she has to correct and enter the letters in the same font again, resulting in the same misreading. , this greatly impedes the speed of character input work.

<Purpose of the Invention> The present invention is intended to solve the above-mentioned problem, and by informing the person writing the characters of the character compression processing process, it eliminates the sense of distrust of the person writing the characters (and speeds up the character input work). The purpose of the present invention is to provide a handwritten character recognition device that achieves this.

<Structure and Effects of the Invention> In order to achieve the above object, the present invention provides an input tablet having a character input area for inputting characters by hand, and an input tablet for inputting characters by hand by bringing a pen tip into contact with the input tablet. A handwritten character recognition device comprising an input pen, a control processing unit that receives handwritten input characters from a character input area of an input tablet and recognizes the characters, and a display unit that displays information related to the input characters or recognition results. , in the control processing unit, means for determining character end points and change points from each stroke of the captured handwritten input character; means for generating a compressed image of the input character configured based on the character end points and change points; The apparatus is provided with means for extracting character features from the image and performing character recognition, and means for sending the compressed image to a display section for display.

According to the handwritten character recognition device of the present invention, since a compressed image of input characters is displayed on the display unit, the person writing the characters can grasp the tendency of the device regarding character recognition processing, and even if there is a misreading as shown in FIG. Even if this occurs, there is no risk of the person filling in the text becoming distrustful of the performance of the device. In addition, since the cause of misreading can be clearly understood by the person writing the text, in the event of misreading, the user can correct and re-enter the text using a font that emphasizes character characteristics, depending on the characteristics of the device.

Figure 14 shows an input image (left figure) in which the number "2" is written with roundness emphasized and its compressed image (right figure).
Since it is closer to the number "2" than ZJ, it is correctly recognized as "2".

Therefore, by performing the character input operation as described above, it is possible to prevent the apparatus from re-reading the characters incorrectly, and the speed of the character input operation can be greatly improved.

<Description of Embodiments> FIG. 1 shows the overall configuration of a handwritten character recognition device according to the present invention.

The illustrated device includes a personal computer main body (hereinafter simply referred to as "computer main body") 1, an input tablet 2
, a printer 3, a floppy disk device 4, etc., and these components are electrically connected by cord lines.

The computer main body 1 is a CRT (Catbode R).
ay Tube) 5 and a keyboard 6, which is equipped with a plurality of function keys for inputting various commands manually. Note that this type of command input section is not limited to the keyboard, but may also be provided on an input tablet (described later) and operated using a pen.

Inside the computer 1 body, there is a CPU (Central CPU).
l Processing Unit), RAM (
Randa*, Access Mea+ory), R
A control processing section including a memory such as OM (Read Only Men+ory) is incorporated, and this control processing section controls each input/output operation and executes a series of character recognition processes including dictionary checking.

The input tablet 2 is for inputting desired characters by hand by bringing the pen tip of a dedicated input pen 9 into contact with the tablet surface and moving it, and the tablet surface includes:
A character input area 7 is provided where the touch of the pen tip can be sensed. This input tablet 2 has a built-in coordinate detection circuit configured with a matrix circuit corresponding to the character input area 7, and this coordinate detection circuit is configured to detect the X coordinate and the Y coordinate of the pen tip that is in contact with the character input area 7. Detect coordinates.

In the character input area 7, a plurality of square-shaped character entry frames 8 are formed in series over multiple stages. These character entry frames 8 are for handwriting input of one character into the frame.After positioning a form on which a similar character entry frame is printed on the tablet surface, insert the input pen into the character entry frame. When characters are handwritten one after another using the CRT 9, each character is recognized one after another, and the recognition results are output to the CRT 5. In this case, in addition to the character recognition result, each time a character is input, an input image and a compressed image of the character are displayed on the screen of the CRT 5.

FIG. 4 shows an example of the screen configuration of the above-mentioned CRT 5. In the illustrated example, there is a display section 5A for displaying a series of recognition results on one screen, and a display section for displaying input images and compressed images, respectively. Sections 5B and 5C are provided.

The input pen 9 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 this, each stroke of the input character from the stroke on the stroke to the stroke on the stroke is obtained as data.

FIG. 2 shows the contents of the memory in the control processing section, and a program area 11 in the figure stores a series of programs for character recognition processing. The work area 12 includes a coordinate storage area 13, a stroke code storage area 14,
It has a square number setting area 15, etc., and a coordinate storage area 13.
indicates the coordinates of each pen position of the input pen 9 moving on the input tablet 2, the stroke code storage area 14 contains the 8-direction code for character strokes (shown in FIG. 3), and the square number setting area 15 contains the characters. Squares required for input (
The number of each character entry frame is stored. Further, the character storage area 16 stores the recognition results for all human characters, and the dictionary areas 17 and 18 store standard data for dictionary checking of alphanumeric characters and kana characters.

FIG. 5 shows the control operation of the control processing section. First, in step 20, the control processing section is initialized and the contents of the work area 12 of the memory are cleared. Next step 2
Step 1 determines whether there is any key input from the keyboard 6 of the computer body 1, and when a key requesting tablet input is pressed, the step 21 determines whether there is a command input or not. The determination of "Is this an input start command?" in step 22 is both "YES" and the process proceeds to step 23. In this case, if another command input key is pressed, step 22 becomes "NO". The process then proceeds to step 41, where the process corresponding to the input command is executed.

In this way, when the input pen 9 contacts the character input area 7 of the input tablet 2 and a predetermined writing pressure is applied, the determination of "Writing?" in step 23 becomes "YES", and the next step 2 is performed.
The number of squares required for writing in step 4 is square number setting area 15.
In the following step 25, the pen position coordinates are detected by the coordinate detection circuit in the input tablet 2.
The coordinates are stored in the memory device/rea 13. Then, when the pen tip moves and a character stroke is drawn within the same character entry frame 8, the determination of "Is it a beautiful handwriting?" in step 26 is "NO".
The judgment of “Is it within the same square?” in step 27 is “YES”
As above, the pen position coordinates are stored in the coordinate storage area 13.
(step 25).

This processing operation is repeatedly executed every time the pen tip position passes a coordinate point, and when the pen tip leaves the input tablet 2 due to the manual completion of one stroke, the determination in step 26 becomes "YES" and the process proceeds to step 30. move on.

In addition, in the process of drawing character strokes, the pen tip is in the character entry frame 8.
If it jumps out, the judgment of “Is it within the same square?” in step 27 will be “NO” and an error sound will be generated in step 28. In this case, the writing of the pen tip from the input tablet 2 will be Therefore, the next step 29 is “YES”.
” and proceed to step 7, step 30.

Step 30 checks whether there is another touch of the input pen within the same character entry frame 8, that is, whether there is a next character stroke for the current input character. ?” is “YES”, step 27
The pen tip movement locus within the same square is continued to be tracked (step 25).

When the input of one character is thus completed and the key instructing the end of the input is pressed on the keyboard 6, the determinations of "Is there a command input?" in step 31 and "Is the input completed?" in step 32 are both "YES". ” and step 3
Proceeding to step 3, the point of change is determined from the character stroke of the input character using the equal division method or the change degree detection method. The equal division method, as shown in FIG. 6, is a method of dividing a character stroke at equal intervals to find change points 53, while the change degree detection method, as shown in FIG. (For example, a point that changes by 90 degrees)
This is a method of detecting the change point 53 as the change point 53. The arrows in Fig. 6 indicate the points at which the stroke is divided, and the arrows in Fig. 7 indicate the points at which the stroke changes by 90 degrees. In both figures, the white circles 51 and 52 indicate the end points of the characters, that is, the writing points and Indicates the point of writing.

After finding the change point of the input character using one of the above methods,
A compressed image (the right-hand diagram in FIG. 6.7) is generated by connecting the change point 53 and character end points 51 and 52 in a straight line, and this compressed image is processed in the next step 33.
It is displayed on the screen of the CRT5. In this case, CRT
It has already been mentioned that, in addition to the character recognition results, the input image of the input character and its compressed image are displayed on the screen of 5 (FIG. 4).

FIGS. 8(1) to (3) illustrate a method of displaying this compressed image, and FIG. 8(1) shows a method of adding change points 53 and character end points 51 and 52 to an input image; FIG. 8(2) shows that changing points 53 and character end points 51 and 52 are added to the input image, and the lines 54 (indicated by broken lines in the figure) that form the compressed image are given a line width and color different from those of the input image. Display method, Figure 8 (3)
is a change point 53 and a character end point 51.5 apart from the input image.
2 and 54 respectively illustrate a method for displaying a compressed image consisting of lines 54 and 54.

The compressed image is displayed using C as in the embodiment shown in FIG.
In addition to using the RT5, as shown in FIGS. 9 and 10, a liquid crystal display 5 provided on the input tablet 2, etc.
5.56 is also possible. FIG. 9 shows a large input image and a small compressed image displayed on a liquid crystal display 55 of, for example, about 3 cm x 3 cm, and FIG. 10 shows a two-stage liquid crystal display 55.
6, immediately next to the recognition result character (see Figure 10 (1)), immediately below (see Figure 10 (2)), or at the end position (see Figure 10 (2)).
(See FIG. 10 (3)) shows the compressed images.

Returning to the flow of FIG. 5, after the compressed image is displayed in step 34, the conversion process of the 8-direction code is performed for each stroke of the human characters (in this case, the character pattern on the compressed image) in step 35. This is executed, and in the following step 36, the converted direction code is stored in the stroke code storage area 14 of the work area 12.

FIG. 11 shows an example of direction code conversion for the character "a", and in the figure, ■■■ indicates the direction code of the character stroke. Note that when performing this code conversion, for example, as shown in FIG. 12, when adjacent direction codes are consecutive (in the illustrated example,
(■■), a so-called rounding process is performed to make the direction code (2) over a long stroke 0-representative as a stroke code.

Next, in step 37, a dictionary matching process is performed on the input characters by referring to the alphanumeric and kana dictionary areas 17 and 18. If the input character matches the standard pattern of any character, the determination in step 38 "Does it match?" becomes "YES" and the process proceeds to step 39, where the recognized character corresponding to the predetermined position on the screen of the CRT 5 is displayed. Is displayed. If the determination in step 38 is aNO'',
Proceeding to step 40, an error message indicating unrecognizability is displayed on the CRT 5.

[Brief explanation of drawings]

FIG. 1 is an external view showing the overall configuration of a handwritten character recognition device according to the present invention, FIG. 2 is a diagram showing a memory map of the device, FIG. 3 is a diagram for explaining an 8-way code, and FIG. 4 is a CRT. FIG. 5 is a flowchart showing the operation of the device; FIGS. 6 and 7 are diagrams for explaining the method of determining the change point; FIGS. 8 (1) to (3) are 9 and 10 (1) to (3) are diagrams showing other embodiments of the display means, and FIG. 11 is a diagram showing the direction code conversion of the character "A". FIG. 12 is a diagram showing an example, and FIG. 12 is a diagram for explaining rounding processing in code conversion. FIGS. 13 and 14 are diagrams for explaining input character compression processing. 1...Computer body 2...Large font 7...Character input area 6...CRT 9...Input pen patent Applicant: Tateishi Electric Co., Ltd. Agent, Patent attorney, Yuki Ki picture],:',,/,:
:ri→/=4 2 pieces of roof/, #2-ta wood body 2, input material 2, and cannibal riff in the shell? , Irukahen-buki / 3 wanderings / 4 Zuken and 2 wandering 3 figures Summer 3 figures together / / figure 72 figures

Claims (5)

[Claims] (1) An input tablet having a character input area for inputting characters by hand, an input pen for handwriting characters by touching the pen tip to the input tablet, and inputting characters by hand from the character input area of the input tablet. The control processing unit includes a control processing unit that captures and recognizes characters, and a display unit that displays information related to input characters or recognition results, and the control processing unit detects character end points and change points from each stroke of the captured handwritten input characters. means for generating a compressed image of an input character configured based on character endpoints and change points; means for extracting character features from the compressed image to perform character recognition; A handwritten character recognition device comprising: means for transmitting and displaying a handwritten character to a computer. (2) The control processing unit is configured with a computer circuit including a CPU and memory.
The handwritten character recognition device described in Section 1. (3) The handwritten character recognition device according to claim 1, wherein the display section is a CRT display device. (4) The handwritten character recognition device according to claim 1, wherein the display section is a liquid crystal display. (5) According to any one of claims 1, 4, and 5, the display unit includes a part that displays an input image of input characters and a part that displays the compressed image. The handwritten character recognition device described.