基于小波矩的印刷体英文字符识别研究

Study of print character recognition based on wavelet moment

基于小波变换思想,将小波矩应用到印刷体英文字符识别当中,通过提取字符图像的小波特征,字符的细节信息得到了完全而稳定的表示,提高了不同类别字符之间的可分性,与传统的几何矩相比,小波矩不但具有一定的几何不变性,抗噪能力也大大地增强.实验结果表明,提出的方法具有较好的识别效果.

Based on the wavelet transform, contribute the wavelet moment to the recognition of the print character, through wavelet feature extraction from character image, the detail information combined in character is expressed entirely and steadily, improves the separability among different type characters. Contrast to traditional geometry moment, wavelet moment has not only geometry fixity, but also enhance the ability of anti-noise. Experiment shows that the method in this paper is effective.