Eigen Space Based Character Recognition Application for Learning Chinese Handwriting in Palembang

Abstract

Eigen space model was proposed in this study for Chinese hadnwriting Recognition. The principle component on image extracted, namely eigen vector and eigen value. These features will be used to recognize each character that written in real time using some kinds of input device. An application system are developed for learning Chinese characters which is designed to be able to recognize hand written characters, pronounciate, and translate it into Indonesian. The system was designed for beginner in Chinese  handwriting learning and only recognize the basic character of Chinese characters. Images of Chinese handwriting used as the input and also a database contained  265 images binary images for feature extraction. The feature extraction and  feature matching algorithm designed based on Eigen Space Model and Euclidean Distance Method. This application used a canvas painted as an input media that made users able to perform a direct input in real time. The system  perform recognition using the proposed method to extract the image of Chinese handwriting with a characteristic value of the tested image and the image of 5 samples per Chinese handriting image data and the average level of Chinese handwriting recognition percentage reached up to 70%. In term of effectiveness, the system is also evaluated by user. The result shows that it is promising tool for learning chinese handwriting