Research And Implementation Of An Online Handwritten Chinese Character Recognition System Based On BiHua

With computer technology, especially the rapid development of Internet technology, mankind entered the information age. In China, many families have personal computers. The use of Chinese characters in computers, first encountered the problem is how to effectively input Chinese characters. Chinese character input directly affecting the computer and information processing application development.Currently, computer keyboard input is the main character input method. English character input into the computer directly, not Chinese which require the use of special technical input. Many Chinese keyboard input methods, mainly boil down to two categories: Pin Yin input method and Codes input method.In the computer to input Chinese characters that have such worries. When using of phonetic input method, user pronunciation is not correct. When using of Codes input method, user do not know Chinese counterparts in the coding. In light of this situation, handwritten input method has become increasingly popular. Written entirely in accordance with the people's daily habits. When writing characters, users only need to input the Chinese characters in"whiteboard"and Computer tracks point to capture handwriting recognition to write Chinese characters, and then completed inputting Chinese characters.This paper focuses on a Chinese character strokes on the on-line handwritten Chinese character recognition system and its realization. The system designed to capture the tracks to identify the strokes of Chinese characters, and then identify the corresponding series of strokes of Chinese characters. Only in the realization of the notebook mouse or touch screen input Chinese characters as a tool, not exclusive"whiteboard."Online handwritten Chinese character recognition can be divided into two broad categories: on the whole word recognition method and based on stroke recognition method. In this paper, it is the latter, Stroke is the first to identify and then find out the strokes of Chinese characters which identify the 2 level of recognition. This feature is based on Handwriting Recognition is a real-time online. In the identification process, stroke recognition is a prerequisite for recognition of Chinese characters and key.This paper uses the strokes identification method which is based on the concept of strokes. Strokes can be further subdivided into strokes of paint. In this paper, eight strokes direction code said.Strokes in the writing process, a series of strokes trajectory of the coordinate point, which is discrete coordinates. Ren computer can distinguish between two adjacent section of the curve is the smallest linear. In this paper, eight strokes direction code labeling each one of the smallest linear respective direction. In this way, each one stroke can produce tracks for writing code sequence that is the direction the current one stroke of the pen sequence. By direction of the sequences filtering, merging after stroke can be the most simple sequences direction. Strokes feature dictionary using dynamic programming with the matching set of all possible strokes, thereby completing the identification strokes, even some of them, including a sum of the parts identification.Due to the randomness of handwriting, handwritten Chinese characters or strokes vary from person to person, and from time to time differences, writing, equipment and environment to another. So in the identification process, the coding sequence is a random direction. This requires that the system is self-learning function that can be used for learning and memory. Many types of machine learning methods, the paper discussed the issue of Stroke Recognition learning grammar.Design based on the Chinese character strokes papers online handwritten Chinese character recognition system in the process, two serial and parallel modes. Whole word recognition discussed five main issues:(1) in Chinese character recognition process for the identification of single stroke after treatment.(2) Continuous Stroke Recognition;(3) Identification of the types of dictionaries and related issues;(4) strings of matching;(5)how to use omitted symbole in input strokes. Finally, also discussed the cell phone input and one way of testing sentence methods.Handwritten Character Recognition System thesis research overhead small savings hardware and software resources can effectively apply to embedded system and microprocessor a great prospect.