基于Gabor-SVM的文字识别方法

摘要

The difficulty and key of Chinese character recognition are feature extraction. In this paper, Chinese character can be treated as an image containing special texture information. Using Gabor filters to extract features, and using SVM method to study these features. The experiment shows that this method performs excellently for Chinese characters feature extraction and recognition.%文字识别的难点和关键在于特征提取,文章把文字版面看作是含有特殊纹理信息的图像,利用G-abor变换,通过纹理分析提取出文字的全局特征.文字特征提取后,对其使用SVM进行训练学习.实验结果表明本方法能够较有效地提取出字符特征,并能有效地对字符进行分类.