KR100356037B1 - Apparatus And Method For Recognition Of Multiple Character In Handwriting Recognition - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

The present invention relates to a handwriting input device capable of recognizing multiple characters and a method thereof.

2. Technical problem to be solved by the invention

According to the present invention, in the case of handwriting input, in addition to being able to write several types of characters in a single handwriting input window several letters, the handwriting input method can reduce the memory while moving the separated handwriting input area flexibly. An object of the present invention is to provide an apparatus and a method for recognizing multiple characters.

3. Summary of the Solution of the Invention

The present invention provides a handwriting input apparatus capable of recognizing multiple characters, comprising: first input means for inputting multiple characters by stroke; Real-time recognition output means for showing the recognition result of the input multi-character; Candidate list output means for showing a candidate list corresponding to the input multicharacter; And output means for recognizing and outputting the input multicharacter.

4. Important uses of the invention

The present invention is used in handwriting input devices and the like.

Description

Apparatus And Method For Recognition Of Multiple Character In Handwriting Recognition}

The present invention relates to a handwriting input device and method for recognizing multiple characters, and more particularly, to a handwriting input device and method capable of easily correcting a mistaken character with a highly accurate recognition ability for various characters. .

Here, multi-letter refers to various kinds of characters such as Korean, English, Japanese, Chinese, and numbers, and at the same time, a word or a combination of words and numbers such as 'terminal', 'Korean', 'handwriting', and 'terminal 026' Say.

Handwriting Recognition in the field of information and communication currently includes a method of inputting letters one by one or a cell method of dividing a predetermined writing input area.

First, the method of inputting one letter at a time is a method in which the input means displays characters corresponding to the input stroke in the input window when the inputter inputs an unlimited stroke anywhere in a single input window without a partition. to be. Hereinafter, this method will be referred to as a " no-compartment method. &Quot; This non-compartment method can be conveniently used without causing any problem in a language arranged in one direction without combining the phonemes, such as English. However, in a language that requires a combination of phonemes for each syllable, such as Hangul, the recognition method and output method for distinguishing and marking the syllables as the inputter intends are complicated. One of the methods is classified into two types, one of which recognizes the strokes input by the inputter as consonants such as consonants or vowels. In this method, the accuracy of character recognition is high because the number of characters to be recognized from the stroke is small, but it can be combined differently from the inputter's intention in the process of combining Korean characters, and the method for overcoming such a combination error is complicated. The other method is to recognize the input strokes input by the input user and display them as Hangul complete letters rather than by phoneme. This method has the disadvantage of increasing the number of characters to be recognized from the stroke, which leads to a large number of recognition errors in the recognition stage and the input method is very complicated and sensitive. Similar to the method of recognizing complete letters in Esau, the number of characters to be recognized one by one and recognized from the stroke is higher than in Hangeul, so the recognition errors and the complexity and sensitivity of the input method become more severe. .

Second, the method of dividing cells after fixing the handwriting input area in advance is to divide the input window into two or more cells, and input the stroke into the appropriate cell according to a predetermined method. Hereinafter, this method will be referred to as a "cell partitioning method." In English, the input window is divided into two cells to distinguish between lowercase and uppercase letters, and in Korean, two or more input windows are used to clarify syllables. The cell is divided into cells, and all characters of the phoneme unit to be combined into one syllable letter are entered in each cell.In Chinese and Japanese, the input window is divided into two or more cells, one in each cell. This type of cell partitioning method is not much inconvenient when applying English input to English input. However, when applying to Korean, Chinese, and Japanese input, it is possible to input single letters without division. Recognition errors can be reduced, but the user must enter only one complete character in a cell Due to the difference with the existing writing habits, the user cannot easily adapt, and thus the input speed may be significantly reduced. Both the no-compartment method and the cell-compartment method are all in one piece, and such advantages and disadvantages are particularly true for all languages. It does not apply equally. In other words, each language has an optimized input method that can facilitate input and increase recognition rate.

On the other hand, various codes and languages such as English, Korean, Chinese, Japanese, numbers, and symbols provide a set of input strokes and recognition algorithms for recognizing each character, and each recognition algorithm is input by a user. The stroke is compared with each of the set of strokes provided in the language and is recognized as a character corresponding to the most similar stroke. However, a part of the strokes forming the stroke set for each language is the stroke forming the stroke set used in the other language. Similar to As a result, when a single input window is mixed for input of multiple languages, input strokes may be mistaken for characters of a language different from the user's intention. Therefore, a single language cannot be input without inputting multiple languages. It is configured to input only, or to input while switching the input language by operating a separate language selection button. When the input language is configured to replace the input language by operating the language selection button, Not only is it troublesome to operate the language selection button, but it is also difficult to configure the input window in a manner suitable for each language.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and is intended to configure the input window of the handwriting input device so that it is possible to input various characters and symbols quickly and easily, and to recognize the input characters with high accuracy. Another object is to efficiently utilize a limited area of an input window configured to input various characters and symbols. Another object of the present invention is to re-enter the entirety of a string when there is an error in a portion of the input string. It is intended to make it easy to correct errors.

1A and 1B are exemplary views illustrating a handwriting input / output window according to the present invention.

2A and 2B are exemplary views when a word is input to a handwriting input window according to the present invention.

3A and 3B are exemplary views when a word is incorrectly recognized when a word is input according to the present invention.

4A and 4B are exemplary views illustrating changes in the input window size according to the present invention.

5A and 5B are exemplary views of an input window when two or more characters are input according to the present invention.

6 is a flowchart of an embodiment of a handwriting input method according to the present invention;

Explanation of symbols on the main parts of the drawings

101: first character input window 102: second character input window

103: third character input window 104: real-time recognition output window

105: candidate list output window 106: output window

107: input window control point 108: candidate list selection button

In order to achieve the above object, the present invention provides a handwriting input apparatus capable of recognizing multiple characters, comprising: first input means for inputting multiple characters by a stroke; Real-time recognition output means for showing the recognition result of the input multi-character; Candidate list output means for showing a candidate list corresponding to the input multicharacter; And output means for recognizing and outputting the input multicharacter.

In addition, the present invention is a handwriting input method applied to a handwriting input device capable of recognizing multi-character, in the case of input by a stroke (stroke) to check the coordinates of the place where the character to be recognized according to the coordinates A first step of determining the type; A second step of displaying a recognition result of the input character; And a third step of outputting the recognition result after the delay time has elapsed.

The present invention is a run-on handwriting input recognition that recognizes even if a user writes a number of letters as a habit of writing in a usual manner. I want to improve.

In addition, it shows the trajectory of the handwriting input and shows the recognized result in real time, so that the recognizer can recognize in real time how to recognize it, and shows the candidate list to increase the recognition rate so that the correct letter can be selected.

That is, the present invention provides a specific character to secure a disadvantage of a method of writing in a fixed, separated box in the input device when a character of various countries or various kinds of characters are inputted to increase the recognition rate of handwriting recognition. The present invention relates to a handwriting input device and a method for inputting a large number of characters by moving the input area in a flexible manner.

The hardware process of basic handwriting recognition is omitted because it is a well-known technique.

Hereinafter, a preferred embodiment according to the present invention will be described in detail with reference to FIGS. 1 to 6.

In particular, in the following description, a stroke refers to a method of inputting by hitting or hitting using a pen or a pen-like tool, as well as a stroke or a stroke. 이중 'and' b 'have double meanings, and it is used as the former or the latter depending on the situation.

1A and 1B are exemplary views illustrating a handwriting input / output window according to the present invention.

As shown in the drawing, the handwriting input window according to the present invention includes a first character (character) input window 101 for inputting Korean characters and a second character (character) input window for inputting numbers ( 102) and a third character (character) input window 103 for inputting an English letter, a real time recognition output window 104 for showing a real time recognition result of the input character, a character being input to increase the recognition rate A candidate list output window 105 for showing a candidate list for the candidate list, a candidate list selection button 108 for selecting any one of the candidate lists, an output window 106 for showing a finally inputted character, and the input It consists of an input window control point 107 for resizing the window by mouth.

That is, FIGS. 1A and 1B illustrate input device units 101, 102, 103, and 107 and output device units 104, 105, 106, and 108 capable of representing strokes with a pen on a touch screen.

2A and 2B are exemplary views when a word is input to a handwriting input window according to the present invention.

That is, in the process of writing the entire word 'terminal' with a pen or a tool similar to a pen in the first character input window 101 on the touch screen, a trace 201 of 'terminal a' is displayed on the real-time recognition output window 104. It shows the recognition result 202 in real time. In other words, if the 'a' is written before the predetermined time elapses after the 'terminal' is written, the 'a' is recognized in real time and the 'terminal a' is displayed on the real-time recognition output window 104. After that, using 'l' will output the result recognized as 'terminal'.

In this case, the result output to the real-time recognition output window 104 can be said to show the zero rank candidate for the character 'terminal' currently input by the user.

That is, if the correct word 'terminal' 212 is recognized in the real-time recognition output window 104 after the word 'terminal' 211 has been input, a predetermined delay time (Time Delay, 0.4 to 0.5 seconds) If no input operation or change is made, the word 'terminal' is output (215).

At this time, the 'terminal' is output to the output window 106 and the candidate list output window 105 has the first rank candidate 'short term' 213 in addition to the 0 rank candidate output to the real time recognition output window 104. Will print.

3A and 3B are exemplary views when a word is incorrectly recognized when a word is input according to the present invention.

If the word 'terminal' 301 is inputted as shown in FIG. 3A, but the word misrecognized as 'terminal 302' appears in the real-time recognition output window 104, the candidate list output window 105 appears in the candidate list output window 105. If a correctly recognized result of the first rank candidate list 'terminal 303' is on the list, after the delay time has elapsed, click 'terminal' 303 with a pen or a similar tool to display the 'terminal' in the output window 106. Output 305.

However, as shown in FIG. 3B, the list is continuously misrecognized as 'short term' 302 in the real-time recognition output window 104 and 'short term' 313 in the first rank candidate list output window 105. If the candidate list is selected after the delay time elapses, pressing the candidate list selection button 108, the four additional candidate list 'terminal' (318), 'dalmanji' (319), 'dalmagi' (320), 'dalmaji' (321) Comes out.

In this case, when the terminal 318 that is a valid candidate list is clicked, the terminal 315 is output to the output window 106.

4A and 4B are exemplary views illustrating a change in an input window size according to the present invention.

When you want to write the word 'terminal mobile phone 026', in the drawing 4a, if you want to write 'terminal' 401 and write 'mobile phones' in succession, but the input window 101 is too small, the input area control point (Writing area control point) ( Clicking 405 and pulling down the first line 403 causes the input window 101 to expand to the second line 404.

That is, when the input window control point 107 in FIG. 1A is clicked and moved downward by a, the end point of the first character input window 101 moves from (x1, y1) to (x1, y1 + a). do.

On the contrary, the area of the third character input window 103 decreases as the first character input window 101 increases.

That is, as shown in FIG. 4B, when the 'terminal' 413 is outputted to the output window 106, the 'cell phone' 411 is written to the first character input window 101 increased by the above process. Through the above-described process, the 'terminal mobile phone' 423 of FIG. 4C is output.

Subsequently, in order to write the number '026', the second character input window 424 may be written over a line as shown in '026' 421 of FIG. 4C.

5A and 5B are exemplary views of an input window when two or more characters are input according to the present invention.

When you want to use 'terminal handset 026' (504) and 'Call' on the next line, you can use 'Enter' (501), 'Enter'. Recognizing the meaning, the cursor 503 moves to the next line 514.

The English letter 'Call' may be written in the third character input window 103. At this time, the recognition of the English letter 'Call' is output through the above process. If the size of the third character input window 103 is small, as described above with reference to FIG. 4, the input area control point 518 is clicked and moved upward 517.

After the process as described above is output to recognize the 'terminal mobile phone 026' (521), 'Call' (522).

6A and 6B are flowcharts of an embodiment of a handwriting input method according to the present invention.

First, the stroke (stroke) trajectory is initialized (601).

When a stroke (stroke) is input on the touch screen with a pen (602), the coordinates of the stroke input position are checked (603) to determine whether the first character input window 101 or the second character input window 102 is used. Alternatively, the type of character to be recognized is determined according to whether the third character input window 103 is used (604).

Next, a 0 rank candidate is output from the recognized result for the stroke input to the real-time recognition output window 104 (605). It is determined whether a predetermined delay time has elapsed (606), and if it has not elapsed, a new stroke (stroke) wait for stroke).

That is, when the word 'terminal' is to be input, the above process is repeated while the term 'term', 'term', and 'group' are input.

In this case, it is determined whether a delay time (0.4 to 0.5 seconds) has elapsed while all of the words 'terminal' have been input (606), and if the delay time is exceeded, the 0 rank candidate among the results recognized for the input word is candidate. As a result, the result displayed in the real time recognition output window 104 is output to the output window 106 (607).

At this time, the candidate list output window 105 outputs a first rank candidate list different from the zero rank candidate displayed on the real-time recognition output window 104 (607).

Next, it is determined whether the output character is the same as the input character (608), if the same, the operation is terminated.

If the output character is not the same as the input character (608), it is determined whether there is a desired list in the candidate list (609) and if one is selected from the candidate list (610). When the candidate list is selected, the previous output result is deleted (611), and after the selected candidate list is output (612), the operation ends.

If there is no desired list in the candidate list (609), the output result is deleted (613) and the process ends.

At this time, if you want to connect and output a new character after the operation is finished, if the input window is insufficient, adjust the input window control point 107 as described in Figure 4 and input a new character.

In addition, when a character line different from a character already input is to be input by changing a line, the character input is performed after the procedure described with reference to FIG. 5. When it is not necessary to change the line, the gesture as shown in FIG. 5 is not necessary. If only the position of the input window is selected, the type of the character to be automatically recognized is determined by determining the coordinates of the input position (604).

In addition, in the present invention, as well as changing the line as described above, it is possible to space, delete the wrong input characters, change, and the like.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the spirit of the present invention for those skilled in the art to which the present invention pertains, and the above-described embodiments and accompanying It is not limited to the drawing.

The present invention as described above can be written as usual writing habits by writing several letters at the same time instead of writing one by one, not only one country letter in one handwriting input window, but several country letters or custom letters and gestures (Delete, Copy, Space, Enter, etc.) can be written at the same time, can be edited by the handwriting input method, the input window to overcome the disadvantage of not writing many characters due to the small size of a single text input window It can be moved flexibly, it shows the trajectory of the letters that the user is inputting, and it can see the result recognized in real time, and even if it recognizes the wrong result, it shows a list of candidates for the recognition result so that the user can select it. Dot, while having the same function as above, it shows high handwritten character recognition rate. And to take up very little memory has excellent effects that can be applied to various fields.

Claims (13)

Handwriting input and output including an input window for inputting strokes by hand, an output window for displaying characters corresponding to the strokes input to the input unit, and recognition and output means for recognizing and outputting the strokes inputted to the input unit as characters or symbols. In the apparatus, The input window is divided into at least two cells, And the recognizing and outputting means is configured to recognize strokes input to each cell of the input window as different types of characters and output them to the output window. Handwriting input and output including an input window for inputting strokes by hand, an output window for displaying characters corresponding to the strokes input to the input unit, and recognition and output means for recognizing and outputting the strokes inputted to the input unit as characters or symbols. In the apparatus, A real-time recognition output window for displaying in real time the character recognized from the stroke entered in the input window; And additionally including a candidate list output window displaying characters similar to the characters displayed in the real-time recognition output window so that the characters displayed in the real-time recognition output window can be replaced when there is an error. If the user selects the characters displayed in the candidate list output window, the recognition and output means outputs the selected characters to the output window, If the user does not select a character displayed in the candidate list output window until a predetermined time elapses after the character is displayed in the real-time recognition output window, the recognition and output means displays the character displayed in the real-time recognition output window in the output window. Handwriting input and output device capable of multi-character recognition, characterized in that configured to output to. The method according to claim 2, The input window is divided into at least two cells, And the recognizing and outputting means is configured to recognize strokes input to each cell of the input window as different types of characters and output them to the output window. The method according to claim 1, A real-time recognition output window for displaying in real time the character recognized from the stroke entered in the input window; And additionally including a candidate list output window displaying characters similar to the characters displayed in the real-time recognition output window so that the characters displayed in the real-time recognition output window can be replaced when there is an error. If the user selects the characters displayed in the candidate list output window, the recognition and output means outputs the selected characters to the output window, If the user does not select a character displayed in the candidate list output window until a predetermined time elapses after the character is displayed in the real-time recognition output window, the recognition and output means displays the character displayed in the real-time recognition output window in the output window. Handwriting input and output device capable of multi-character recognition, characterized in that configured to output to. The method according to claim 1, The output window is divided into three cells, The first cell is used as a real-time recognition output window for displaying the characters recognized from the stroke entered in the input window in real time, The second cell is used as a candidate list output window displaying characters similar to the characters displayed in the first cell so that the characters displayed in the first cell can be replaced when there is an error. When the user selects the text displayed in the second cell, the recognition and output means outputs the selected text to the third cell. If the user does not select a character displayed in the second cell until a predetermined time elapses after the character is displayed in the first cell, the recognition and output means outputs the character displayed in the first cell to the third cell. Handwriting I / O device capable of multi-character recognition, characterized in that configured to. The method according to any one of claims 1 and 3 and 5, A handwriting input / output device capable of multi-character recognition, characterized in that the boundary between the cell partitioning the input window and the cell is configured to be movable. The method according to claim 6, The input window is divided into three cells, The recognition and output means recognizes and outputs the stroke input to the first cell of the input window in Korean, recognizes and outputs the stroke input to the second cell as a number, and translates the stroke input to the third cell into English and Japanese. And a multi-character recognition capable input / output device characterized in that it is configured to recognize and output any one type of character selected from the multi-character group consisting of Chinese characters. In the handwriting input method applied to a handwriting input device that can recognize multiple characters, A first step of determining a type of a character to be recognized according to the coordinates by checking the coordinates of the stroked place when there is an input by a stroke; A second step of displaying a recognition result of the input character; And A third step of outputting the recognition result after the delay time has elapsed Handwriting input method comprising a. The method of claim 8, A fourth step of outputting a candidate list corresponding to the input character; Handwriting input method further comprising. The method of claim 9, The third step, A fifth step of searching for the candidate list if the recognition result output after the delay time is different from the input character; And A sixth step of outputting the list if the desired list is found as a result of the search; Handwriting input method comprising a. The method of claim 8, The coordinates of the first step, Handwriting input method characterized in that the range can be modified according to the user's adjustment. The method according to claim 6, And an input window control point used to move a boundary between the cell partitioning the input window and a cell. The second cell of the input window is disposed between the first cell and the third cell, When the user pulls the input window control point, the boundary between the first cell and the second cell and the boundary between the second cell and the third cell are configured to move in the same direction. I / O device. The method according to any one of claims 2 and 4 and 5, And the candidate list output window comprises a touch screen for selecting a character by touching it with a handwriting pen or a finger.