KR100297815B1 - System for recognizing korean written characters - Google Patents

Abstract

PURPOSE: A system for recognizing the Korean written characters is provided to rapidly and accurately recognize the Korean written characters by limiting a format of written characters corresponding to the characteristic of Korean. CONSTITUTION: An input device(100) is equipped with an electronic pen, and a user generates a stream of X-Y coordinates according to the locus written by the electronic pen. A microprocessor(104) receives the output of input device, recognizes the inputted Korean characters, and outputs the result as a character code to an output device(112). The microprocessor loads a character combination auto motor(106) and a stroke recognizer(108) in order to recognize the Korean characters. A ROM stores a character recognition process program and various reference data for character recognition process. A RAM temporary stores the data according to the character recognition process. The output device outputs the result of character recognition process by enabling the user to see the result.

Description

Hangul Cursive Character Recognition System

The present invention relates to a system for recognizing handwritten characters, and more particularly to a system for recognizing Hangul handwritten characters.

In recent years, the use of personal information terminal devices such as PDA (Personal Digital Assistant or Personal Digital Accessory), HPC (Hand-held Personal Computer), and the like are increasing rapidly. These personal information terminal devices have tended to become smaller and lighter in personal computers. Accordingly, the necessity of a handwriting input method has emerged as an input method that replaces a bulky keyboard. However, in the handwriting input method, a problem of accuracy acts as an obstacle to practical use, and inconveniences to a user's input speed.

In particular, the problem with the Korean handwritten character recognition method is that it is difficult to find the boundary between the phonemes. That is, when more than ten strokes are input from one stroke or two strokes, the recognition speed is naturally slowed down, and the writing speed is naturally lowered as the recognition must be considered in consideration of all possible combinations generated therefrom. In addition, in this case, ambiguous characters frequently occur, which causes a drop in accuracy.

Therefore, the necessity of a method of limiting a user's handwriting form has been discussed to improve the accuracy and speed of recognition. For example, in the English-speaking world, a technique for dramatically increasing the accuracy of recognition and the writing speed is proposed by writing all the alphabets (alphabet) in one stroke under the name of unistroke. This is US Patent No. 5,596,656, published by David Goldberg, issued January 21, 1997, "UNISTROKES FOR COMPUTERIZED INTERPRETATION OF HANDWRITING." This method is now widely used in PDAs.

FIG. 1 illustrates the form and handwriting order of the unistroke alphabet symbols proposed in Patent No. 5,596,656, including numbers and symbols. 1 shows each of the unistroke symbols as arrows having a unique shape, the writing sequence proceeds from the start point to the end point of the arrow. In this way, since one stroke represents one alphabet, the recognition result is output on the screen immediately after the user writes one stroke without combining strokes. Therefore, since the shape is designed so that the distinction by angle between the alphabets can be made reliably, it can be recognized accurately.

However, the unistroke system developed in the English-speaking language as described above is suitable for the alphabet to be interpreted due to the characteristics of the character, but it cannot be applied to recognize Korean handwritten characters. That is, because Hangul has many kinds of phonemes, complicated forms, and writing together, there are many differences between the two, so it could not be applied as it is.

As described above, the Hangul handwritten characters have difficulty in finding the boundary between the phonemes, thereby reducing the accuracy and speed of recognition. On the other hand, although there is a unistroke system that can improve the accuracy and writing speed of recognition, this is not only suitable for English-speaking characters, but also has a problem that can not be applied to Hangul having different character characteristics.

Accordingly, an object of the present invention is to provide a system that can accurately and quickly recognize Hangul handwritten characters by restricting the handwriting form according to the characteristics of Hangul.

1 is a view showing the form and the writing order of the unistroke alphabet symbols,

2 is a view showing the form and the writing order of the basic phonemes and repeating codes defined according to an embodiment of the present invention;

3 is a block diagram of a system for recognizing a Korean handwritten character according to an embodiment of the present invention;

4 is a state transition diagram of the character combination automer of FIG.

According to the present invention for achieving the above object, each basic phoneme including a terminal sound and a vertical and horizontal vowels is written in one stroke with a Hangul Unistroke symbol defining a unique form and a writing order, and the remaining compound phonemes are Hangul Unistroke. Combining the symbols are characterized in that the writing on each of two strokes.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description and the annexed drawings, numerous specific details are set forth in order to provide a more thorough understanding of the present invention, such as the form, writing order, and processing details thereof. It will be apparent to those skilled in the art that the present invention may be practiced without these specific details. And a detailed description of known functions and configurations that may unnecessarily obscure the subject matter of the present invention will be omitted.

First of all, the present invention allows the concept of unistroke as described above to be applied to a Korean handwritten character recognition system, and a user can more accurately use a digitizer or a touch-sensitive screen panel without any inconvenience. Hangul handwriting form that can be input by hand and a recognition system using the same.

2 is a diagram illustrating the form and the writing order of basic phonemes and repetitive codes defined according to an exemplary embodiment of the present invention. In FIG. 2, "o" represents a starting point position of a stroke. As described above, since Hangul has a larger number of phonemes and complicated forms than English, it is difficult for a user to memorize and write the English Unistroke system as it is. Therefore, it is necessary to redefine the concept of unistroke to suit Hangul. In the present invention, the shape of the Hangul phoneme is defined as shown in FIG. 2 under the following principle.

First, in the present invention, as shown in Table 1, Korean phonemes are divided into basic phonemes and compound phonemes. In the present invention, the basic phoneme shown in Table 1 is defined to be written in one stroke, and the composite phoneme is written in two strokes as the sum of the basic phonemes.

Kind of Jaso Basic phoneme Consonant ㄱ ㄴ ㄷ ㄹ ㅁ ㅂ ㅇ ㅋ ㅋ ㅋ ㅋ ㅋ collection ㅜ ㅐ ㅑ ㅒ ㅓ ㅔ ㅕ ㅖ ㅗ ㅗ ㅠ ㅡ ㅣ Compound A double consonant ㄲ ㄸ ㅃ ㅆ ㅉ Lucky consonants ㄳ ㄵ ㄶ ㄺ ㄻ ㄼ ㄽ ㄾ ㄿ ㅀ ㅄ Double Vowels ㅘ ㅙ ㅚ ㅝ ㅞ ㅟ ㅢ

Here, basic phoneme means any phoneme written on the Korean Hangul keyboard for computer except double consonants such as ㄲ, ㄸ, ㅆ, ㅉ ㅃ. Each basic phoneme corresponds to the same concept as the English unstroke alphabet and is represented by one stroke. This is called "Hangul Unistroke" in the present invention. A compound phoneme is defined as the sum of two consonants in the case of consonants and the sum of the horizontal and vertical vowels in the case of vowels. Therefore, all compound phonemes should be written in two strokes. However, in the case of double consonants, when the sum of the basic phonemes is expressed, the ambiguity that can be interpreted as 'each' or 'near' occurs when the 'ㄱㅏ ㄱㄱㅏ' is written. The principles defined in this way simplify the shape of the phoneme, making it possible to design a handwritten phoneme shape close to the original shape of the phoneme, thus making it easier for a user to memorize the shape of the handwritten phoneme. It also has the advantage that a character who is familiar with the existing keyboard can input characters in the same combination as tapping the keyboard. In order to maintain this advantage, a character combination automer is used as shown in FIG. 4, in which the Hangul character combination automata is partially modified.

In defining the basic phoneme shape of FIG. 2, it is considered to make it easy to distinguish when recognizing while maintaining the original form of the Hangul phoneme. Hangul consists of 51 consonants with 30 consonants and 21 vowels, which are combined from 24 basic forms. Hangul scribes are also used to writing different phonemes by combining basic forms. Therefore, the phoneme shape is defined so that the combined phonemes can be easily associated with the basic form. The present invention allows the user to freely write or write continuously in accordance with the defined phoneme shape. Phoneme designed based on this principle is shown in FIG.

The handwritten phoneme shape designed in FIG. 2 maintains a shape similar to the real phoneme in the case of consonants, and is easily associated with the actual phoneme shape. In the case of vowels, 'ㅏ', 'ㅓ', 'ㅗ', 'TT' and 'ㅑ', 'ㅕ', 'ㅛ' and 'ㅠ' are designed by symmetrical and rotating the same shape. It is designed to be easy for users to memorize by linking 'ㅒ', 'ㅔ' and 'ㅖ'.

Figure 3 shows a block diagram of a system for recognizing Hangul handwritten characters in accordance with an embodiment of the present invention. The input device 100 includes an electronic pen 102, and generates a stream of x, y coordinate points by a two-dimensional coordinate system according to a trajectory of a user writing using the electronic pen 102. As the input device 100, a digitizer or a touch screen panel is used as described above. The microprocessor 104 inputs the output of the input apparatus 100 to recognize the Korean handwritten characters inputted as shown in FIG. 2 and outputs the recognition results to the output apparatus 112 as character codes. In the microprocessor 104, a letter combination automerter 106 and a stroke recognition unit 108 are loaded for Korean handwritten character recognition. The memory 110 includes a read only memory (ROM) and a random access memory (RAM). The ROM stores a character recognition program of the microprocessor 104 and various reference data for character recognition processing according to an embodiment of the present invention. The RAM temporarily stores data according to the character recognition process of the microprocessor 104. The output device 112 may be, for example, a liquid crystal display device, and outputs a character recognition result by the microprocessor 104 so that a user can read it.

In the system of FIG. 3, the character combination automerter 106 receives a coordinate point sequence of a character pattern written by the user whenever the user writes using the input device 100. The character pattern is made up of a set of strokes, and one stroke is a sequence of ordered coordinate points representing a trajectory of movement of the electronic pen 102 from its contact with the surface of the input device 100 to its fall from the surface. The input apparatus 100 continuously tracks the user's handwriting and transmits the coordinate point sequence to the character combination automerter 106 after receiving one handwriting, and continuously receives the input.

The letter combination automerter 106 is represented by the state transition diagram of FIG. That is, whenever a stroke is input from the input device 100, the stroke recognition unit 108 is called to receive a stroke recognition result to generate a character code. One character is composed of a minimum of two strokes and a maximum of six strokes by the character combination automer 106 as shown in FIG. When inputting a writing type phoneme designed using the letter combination automerter 106, the user can continuously input characters regardless of the writing or the writing. That is, in the conventional Korean recognition, only one character is input or recognized, or the character is separated and recognized by an arbitrary method. However, in the present invention, the phoneme can be continuously input like the keyboard.

The stroke recognition unit 108 is a process of mapping a stroke written by a user to one of phonemes. The stroke recognition unit 108 of the present invention may use any conventionally known character recognition method. In other words, the present invention is independent of the method for recognizing stroke. However, since the form is defined to make it easy to distinguish between elements, it is possible to recognize strokes quickly and accurately even with a simple method. Accordingly, in the present invention, regardless of the method of recognizing a stroke, a system for defining a form of a phoneme to be recognized and combining the same to generate a character is defined. When the stroke is recognized by the stroke recognition unit 108, the stroke recognition result is transmitted to the character combination automer 106 again, and when one character is completed, the character code is output to the output device 112 to recognize the character. It is over. This process is repeated until the user stops writing.

Therefore, as the Korean handwriting character is recognized by applying the concept of unistroke in Hangul, the handwritten character can be almost completely recognized, and it is possible to continuously write fast. In addition, since the amount of memory used by the recognizer is reduced, the burden on the storage of a small computer can be reduced.

Meanwhile, in the above description of the present invention, specific embodiments have been described, but various modifications can be made without departing from the scope of the present invention. In particular, unlike in the embodiment of the present invention, the type and writing method of the phoneme to be written in two strokes may be modified as necessary. For example, instead of using a repetition symbol in the expression of the double consonants, the double consonants can be expressed in a single stroke, and the vowels 'ㅐ', 'ㅒ', 'ㅔ', and 'ㅖ' can be modified to be written in two strokes. In addition, any number of modifications or changes may be made to some or all of the phoneme forms defined in the embodiments of the present invention. Therefore, the scope of the invention should not be defined by the described embodiments, but should be defined by the equivalent of claims and claims.

As described above, the present invention can remarkably improve the accuracy and speed of recognition by applying the concept of unistroke in Korean as well and enable fast writing continuously. In addition, the smaller amount of memory used by the recognizer reduces the burden on the memory of microcomputers, helping to compete in price.

Claims (6)

In the system for recognizing Hangul handwritten characters, A compound phoneme defined to write on two strokes by combining the Hangul unistroke symbol that defines its own form and handwriting order to write in one stroke for each basic phoneme including terminal sound and vertical and horizontal vowels. An input device for generating a coordinate point sequence by a two-dimensional coordinate system according to a trajectory that is selectively combined and written by a user, A stroke recognizing unit for recognizing a stroke of each of the written symbols by corresponding to one of phonemes by inputting the coordinate point sequence; And a character combination automator for recognizing the written Hangul characters and generating a character code corresponding to the recognized phonemes by the character combinations according to the handwriting type. The Hangul handwritten character recognition system according to claim 1, wherein the character combination automator comprises one character in a minimum of two strokes and a maximum of six strokes. The Hangul handwritten character recognition system according to claim 1, wherein the input device is a digitizer or a touch screen panel. In the system for recognizing Hangul handwritten characters, Hangul Unistroke symbol that defines its own shape and handwriting order to write in one stroke for each basic phoneme including terminal sound and vertical and horizontal vowels, and a compound character defined to write in two strokes by combining the Hangul Unistroke symbols Selected by the user is a combination of the small consonants and the double consonants and the repetitive symbols defining a unique shape and writing order to write in two strokes in combination with the Hangul unistroke symbol for each of the compound consonants in the compound phoneme An input device for generating a coordinate point sequence by a two-dimensional coordinate system according to the written trajectory; A stroke recognizing unit for recognizing a stroke of each of the written symbols by corresponding to one of phonemes by inputting the coordinate point sequence; And a character combination automator for recognizing the written Hangul characters and generating a character code corresponding to the recognized phonemes by the character combinations according to the handwriting type. 5. The Hangul handwritten character recognition system according to claim 4, wherein the character combination automator comprises one character from at least two to at most six. The Hangul handwritten character recognition system according to claim 4, wherein the input device is a digitizer or a touch screen panel.

Images