关键词: 图像汉字, 小波变换, 边缘检测, 特征提取, 笔段合成

Abstract: Based on the cognitive view of the bionic pattern recognition, from the structure mechanism of Chinese characters and human’s habits of understanding it, this paper puts forward an image Chinese characters recognition method based on wavelet transform. The specific rules about image Chinese character strokes features extraction are developed. Image Chinese characters outline and strokes outline are detected with wavelet transformation. The effective strokes information is extracted and combined. Strokes or a Chinese character is generated. The simulation results prove this method can improve the accuracy and stability of the image Chinese characters stroke features extraction. It has higher recognition rate for print and normative handwritten image Chinese character.

Key words: image Chinese character, wavelet transform, edge detection, features extraction, strokes combine