JPS61138383A - Recognizer of hand-written character - Google Patents

Abstract

PURPOSE:To attain an input capable of discriminating a small character and a capital by providing a rewriting means which detect a character entry frame through touch with an input pen, stores a character conversion table with a set composed of a capital and a small character and fetches an opposite one for pairing with the set. CONSTITUTION:When the input pen 7 contacts the character input area 8 of an input tablet 2 to apply the prescribed pen pressure, a timer for detecting the termination of one character input starts. The lattice number of the character entry frame 10 is stored, and coordinates at a pen position on the character entry frame 10 are written. Upon the completion of one character input,a character recognizing subroutine is called to extract character features, which are collated with standard data stored in a dictionary area. When the result is equal to any of standard patterns, the recognized character is displayed on a CRT5, and simultaneously set in a recognized result memory area. When a small character is converted into a capital, the tip of the input pen 7 is brought into contact with a small character key among mode specifying keys 9 and the character frame 10 where a top character to be converted is positioned.

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> The handwritten character recognition device of the present invention stores in advance a plurality of pairs of uppercase and lowercase letters in a memory as a character conversion table, and when inputting characters, a character entry frame to be converted is specified. The system is configured to refer to the character conversion table, extract the partner character that pairs with the character to be converted, and rewrite the contents of the specified character entry frame, thereby simplifying character input and correction work. and speeding up the process.

<Background of the Invention> In a typical handwritten character recognition device, when characters are handwritten and input into an input tablet, character recognition processing is executed for each character, and the recognition results are output to a display device or the like. By the way, some of these input characters include, for example, the kana characters "ya" and "ya".
, ``Yu'' and ``Yu'', ``Yo'' and ``Yo'', ``Tsu'' and ``Tsu'', which are the same characters but are case-sensitive. If there is a type of character input, it is necessary for the device to clearly distinguish between the two and recognize them. In the case of this kind of kana characters, the uppercase and lowercase letters have the same font, only the size and position are different, so if the device was equipped with a function to determine the area occupied by the character and its position, It is possible to recognize lowercase letters and therefore distinguish between uppercase and lowercase letters, and there is no need to prepare a separate dictionary for lowercase letter matching.

However, when it comes to English letters (alphabet), uppercase and lowercase letters have completely different fonts, so in order to distinguish lowercase letters from uppercase letters and recognize them, it is necessary to prepare a dictionary for lowercase letter matching, which requires an increase in dictionary capacity and This results in higher costs. However, lowercase English letters are more rounded than angular uppercase letters, such as "a", rdJ, and rnJ.
and rhJ, and characters such as "r" and "v" are likely to be mixed, and this can cause misreading or unrecognizability. When this misreading occurs, it is necessary to re-enter the characters, and in this case, it is necessary to take measures such as emphasizing the characteristics of the re-entered characters in order to clearly distinguish them from the misread characters. This makes the input operation complicated and limits the speed of the operation. Furthermore, as a countermeasure against misreading, there is a method in which a touch-type menu board is separately provided, but in this case, space is required to provide the menu board on the input tablet, which has disadvantages such as reducing the effective space for writing characters.

<Object of the invention> This invention is intended to solve the above problems all at once.
It is an object of the present invention to provide a handwritten character recognition device that can input and recognize lowercase letters in English letters, etc. while clearly distinguishing them from uppercase letters.

<Configuration and Effects of the Invention> In order to achieve the above object, the present invention provides a control processing unit that executes character recognition processing with a character entry frame in which a character to be converted specified by contact with an input ben is written. means for detecting, means for storing a character conversion table having a plurality of sets of uppercase and lowercase letters, and a partner character that pairs with the character to be converted by referring to the character conversion table based on the specification of the character entry frame. The system is equipped with a means to extract the text and rewrite the contents of the specified character entry box.

According to the present invention, it is possible to input lowercase letters, such as English letters, while distinguishing them from uppercase letters by only using a point-touch operation of the input pen, thereby improving the simplicity and speed of the operation. In addition, since there is no need to prepare a dictionary for lowercase letter matching, it is possible to save dictionary memory capacity and reduce costs, and since there is no need to perform dictionary matching for lowercase letters, the matching processing time can be reduced.
It is possible to speed up recognition processing. Furthermore, since there is no risk of misreading, there is no need to provide a bench-touch menu board or the like as a countermeasure against misreading, and the effective space for writing text is increased, achieving the remarkable effects of achieving the purpose of the invention.

<Description of Embodiments> FIG. 4 shows the overall configuration of a handwritten character recognition device according to the present invention.

The illustrated device includes a personal computer main body 1 (hereinafter simply referred to as "computer main body"), 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 (Cathode Ra).
y Tube) 5 and a keyboard 6, and the keyboard 6 is provided with a plurality of function keys for manually inputting various commands. Inside the computer main body 1, there is a CPU (Central Proc.
In addition to RAM (Rand
om Access Memory), ROM(Re
A control processing unit including memory such as +adi+:1QaqMemory) is incorporated, and this control processing unit controls input/output operations in each input/output unit, and executes a series of character recognition processes including dictionary matching. Ru.

The input tablet 2 has a character input area 8 for handwriting inputting characters by touching and moving the input pen 7 on the upper surface thereof, and a character input area 8 for handwriting inputting characters by touching the input pen 7 at a point and inputting character types (alphanumeric characters, numbers, kana characters, A plurality of mode specification keys 9 are provided for specifying kanji (kanji, symbols, etc.) and lowercase letters, and the character input area 8 has a plurality of vertical and horizontal squares for entering a large number of characters for each character. The entry frame 10 is formed into sections.

The input pen 7 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 or a form placed on the input tablet 2 and a predetermined writing pressure is applied. Therefore, input tablet 2
By detecting the moving locus of these lines of magnetic force, each stroke of the input character, from handwriting to handwriting, is obtained as data.

FIG. 5 shows the contents of the memory in the control processing section,
A program area 11 in the figure stores a series of programs for character recognition. The work area 12 has a coordinate storage area 13, a stroke code area 14, a square number setting area 15, a conversion square number setting area 16, a lower case mode storage area 17, a recognition result storage area 18, a candidate character area 19, and coordinate storage. The area 13 contains the pen position coordinates of the input pen 7 on the input tablet 2, the stroke code area 14 contains the 8-way code for character strokes, and the square number setting area 15 contains the squares for character input (character entry The conversion square number setting area 16 is filled with the character to be converted.The square number specified with the input pen 7 is written in the lowercase character mode storage area 17. In the lower case mode, the recognition result storage area 18 stores character recognition result data for the current input character, and the candidate character area 19 stores candidate characters for the current input character.

Further, the character storage area 20 stores recognition results for all input characters, and the dictionary area 21 stores mark data for dictionary comparison of alphabetic characters, numbers, and kana characters.

Furthermore, the character conversion table area 22 stores a character conversion table that restores sets of uppercase and lowercase letters for all English characters. For example, in the illustrated area 23, a capital letter rAJ and its lowercase letter rfaJ are set, in the subsequent area 24, a capital letter rBJ and its lowercase letter "b" are set, and in the same way, uppercase and lowercase English letters form pairs. are set sequentially for each group.

Although the present embodiment has only a character conversion table for English characters, character conversion tables for other characters (for example, kana characters) may also be prepared.

FIG. 1 shows the operational processing flow of the control processing section. First, in step 30, a stack pointer is initialized for the memory of the control processing section, and the contents of the work area 12 are cleared. Next, in step 31, it is determined whether a command has been input from the keyboard 6 of the computer main body 1, and when the key requesting tablet input is pressed, the determination in step 31 as to "Is there a command input?" And, the judgment of “Start input?” in step 32 is “Y”.
ES", and the process proceeds to step 33. Then, when the input pen 7 contacts the character input area 8 of the input tablet 2 and a predetermined writing pressure is applied, the determination of "writing" in step 33 becomes "Y".
ES'', and in the next step 34, a timer that detects the end of one character input is started.

This timer is included in the CPU of the control processing section, and in the case of this embodiment is set to about 2 seconds, for example. In the next step 35, it is determined whether or not writing is an input operation for the first character entry frame 10. If the determination is "YES", the process advances to step 36, and the square number of the character entry frame 10 is changed to the square number. Store it in the number setting area 15 and proceed to the next step 3.
At step 7, the coordinates of the pen position within the character entry frame 10 are written into the coordinate storage area 13.

Then, when the pen tip moves and a character stroke is drawn within the same square, the judgment of ``Is it a good handwriting'' in step 38 is NO?
, the determination of “Is it within the same square” in the following step 39 is “YES”.
S''', and the coordinates of the constituent points of the character stroke are sequentially written into the coordinate storage area 13. In this case, if the character stroke protrudes outside the frame, step 39 is "No", and step 40 is performed. An error sound is generated. When the pen tip leaves the input tablet 2, the determination of "Is it a Japanese handwriting?" in step 38 or step 41' becomes "YES", and further in steps 42 and 43, the pen tip leaves the input tablet 2. It is checked whether there is another stroke of the input pen 7, that is, whether the next character stroke exists.

When the input of one character is completed and the one-character input end detection timer times out, the determination in step 44 becomes ``YES'' and the timer stops clocking in step 47.On the other hand, at this stage, if the keyboard 6 When the input end key is pressed, both the determination of "Is there a command input?" in step 45 and the determination of "input completed?" in step 46 become "YES''', and the process proceeds to step 47 as described above. move on.

Once you have completed entering one character as above, proceed to the next step 4.
At step 8, the character recognition subroutine shown in FIG. 2 is called, and the process moves to character recognition processing. First, in step 49 of FIG. 2, character features are extracted from the input characters, and in the following step 50, these character features are compared with standard data stored in the dictionary area 21. If the result matches any of the heading patterns, the judgment in step 51 is "=
YES", and the recognized characters are displayed on the CRT 5 in step 52, and the recognition result data is set in the recognition result storage area 18 in step 53. If the judgment result in step 51 "Is it a match?" is NO. ”, the process advances to step 54, where an error message indicating unrecognizability is displayed on the CRT 5.If a key operation to specify the end of the character input operation is performed at this stage, the process returns to the main routine and step 55 is performed. °'YES', the system waits for the next command input, but if the next character is to be input, the determination at step 55 is °'NO''', and the process returns to step 33 to input the next character. to wait.

After the initialization in step 30, if a command other than the tablet input command is input, the determination in step 32 becomes 'No' and the process proceeds to step 56, where the command processing corresponding to the command input is executed. .

Now repeat the above steps and for example rAJrLJll
When each alphabetic character of JrHJrAJ is input by hand in sequence,
On the screen of the CRT 5, as shown in Figure 6 (1), the English characters [ALPHAJ] are displayed, and the cursor 2
5 will wait at the character position next to rAJ at the end.

Next, when converting the above English letter "ALPHAJ" to a lowercase letter, the person writing the character should touch the lowercase key of the mode designation keys 9 with the pen tip of the input pen 7, and position the first character rAJ to be converted. The input pen 7 is brought into contact with the character entry frame 10 to be written.As a result, the determination in step 33, "Is this writing?" becomes "YES", and the determination in step 35, "Is this the first input?"teeth"
Since the answer is "No", the process proceeds to step 57. In step 57, it is determined whether the square number of the character entry frame 10 that the pen touched is greater than the square number of the character input immediately before ("A" in this case). , if this judgment is ``YES'', the process moves to the normal character input process, but in this case, this judgment becomes ``No'', and the process moves to the lowercase character conversion process of step 58, and the cursor position is also changed. It returns to the position of the first character "A" (as shown in FIG. 6(2)).

FIG. 3 shows a flowchart of a subroutine for converting lowercase letters. When this subroutine is called in step 58, first, in step 70.71,
It is determined whether or not the lowercase key of the mode designation key 9 has been operated by pen contact. In this case, in step 70 "
Is it a command input cloth? ” and step 71 “
Is it a lowercase designated cloth? ” are both “YES”, and in step 72, the lowercase mode is set in the lowercase mode storage area 17, and the next step 73 is set.
In this step, the square number related to the conversion is set in the conversion square number setting area 16.

If the next character input operation is resumed without inputting the input start command at this stage, step 70 is "NO", step 74 is "YES", and an error sound is generated (step 75). In addition, when another command is input from the keyboard 6, step 70 is ``YES''.
", step 71 is °' N O" and the command instructs to end the input, step 76 is "YES" and returns to the main routine, and if the command is otherwise If step 76 is “NO”
The process then proceeds to step 77, where command processing corresponding to the command input is executed.

Thus, in step 78, the square number setting area 1
Based on the difference between the contents of 5 and the contents of the conversion square number setting area 16, the character required for conversion (in this case, the first character "A"
) is stored in the character storage area 20, in the next step 79, the character set at that address is retrieved, the character conversion table is searched, and the character to be retrieved is searched. :lzono1 chi. And when that character is detected in area 23,
The lowercase letter 1''aJ forming a pair with this is read from the character conversion table and set at the address in the character storage area 20 (step 80).

FIG. 6(3) shows the display screen of the CRT 5 at this stage, with the first character replaced with a lowercase "3".
Furthermore, the cursor 25 has moved to the position of the next character "L".

When the conversion of one character to lower case is completed as described above, in the next step 81 and 82, it is determined whether or not the next handwriting has been written and whether a command has been input. If Step 81 is “Drawing?
” is “YES”, the cell number of the character entry frame 10 related to the writing is checked in step 83, and it is determined whether the writing is for character input or for character conversion. judge. And step 83 is “
When the determination is ``YES'', the input is continued and the process returns to the main routine, and after the determination of ``NQ'' in step 59, the process moves to step 36.

On the other hand, when the determination in step 83 is "No", the process moves to the next lower case conversion procedure, returns to step 73, and sets the square number of the character entry frame 10 related to the conversion. No operation required.

In FIG. 1, if there is a command input immediately after a character input operation and before the time-out in step 44, if the command input is related to a lowercase key operation, step 45 is ``YES'' and step 46 is ``YES''. °'No",
YES in step 60, and after the character recognition process shown in FIG. 2 is executed for the immediately preceding input character in step 61, the process proceeds to step 62, a lowercase character conversion process.

When the subroutine for converting lowercase letters shown in FIG. 3 is called at step 62, the lowercase mode is first set in the lowercase mode storage area 17 at step 84, and then the square number for conversion is specified at step 81. Wait for a while, and it will start converting to lower case. And this step 62
After the lowercase character conversion process is completed, it is checked in step 65 whether or not the input is completed, and if the determination is "NO", the process returns to step 36 and the character input process is restarted.

On the other hand, if the above-mentioned command input has contents other than lowercase letter designation, step 60 becomes ``NO'''', and after step 63 similarly executes the character recognition process for the previous input character, step 64 inputs Various command processes are executed according to the command.

Thus, similar lower case conversion processing is performed as rAJ r L jr
When executed for PJrHJrAJ, the converted characters [aJ
rlJrpJrhJraJ is displayed, and the power input 25 has moved to the position next to the last converted character "a".

In the above embodiment, when converting lowercase letters, it is essential to specify the lowercase mode using the mode specifying key 9. However, the present invention is not limited to this.
It is also possible to automatically shift to lowercase character conversion processing only by specifying 0. In this case, in order to distinguish it from processing such as character correction, it is necessary to require operation of a character correction key or the like for other processes.

[Brief explanation of drawings]

1 is a flowchart showing the operation of the apparatus shown in FIG. 4 according to the present invention, FIG. 2 is a flowchart showing the character recognition subroutine of FIG. 1, and FIG. 3 is a flowchart showing the lowercase conversion subroutine of FIG. 1. Flowchart, FIG. 4 is an external explanatory diagram showing a configuration example of a handwritten character recognition device according to the present invention, FIG. 5 is a diagram showing a memory map of the device shown in FIG. 4, and FIG. 6 is a diagram showing an example of display on a CRT. It is.

Claims (3)

[Claims] (1) An input tablet with multiple character entry frames set in an area for inputting characters by hand; an input pen for inputting characters by hand by touching the pen tip to the input tablet; and an input tablet for inputting characters by hand. and a control processing section that reads characters from an input tablet and executes character recognition processing, and the control processing section detects a character entry frame in which a character to be converted specified by contact with an input pen is written. , means for storing a character conversion table having a plurality of sets of uppercase and lowercase letters;
A handwritten character recognition device comprising means for referring to a character conversion table based on the specification of the character entry frame, extracting a partner character that is paired with a character to be converted, and rewriting the contents of the specified character entry frame. . (2) The handwritten character recognition device according to claim 1, wherein the character conversion table is a table for converting uppercase letters to lowercase letters for English characters. (3) The input tablet is provided with a mode designation key that designates a lowercase letter by contact with an input pen, and the control processing unit is provided with means for determining mode designation of a lowercase character,
The handwritten character recognition device according to claim 1, wherein the handwritten character recognition device is configured to convert the characters in the specified character entry frame to lower case based on the determination result.