摘要
利用小波分解提取文字的结构特征,根据文字的大体结构分类汉字,将汉字的分形维作为特征向量,以识别汉字。利用已有数据,在小范围内,验证由分形维识别英文字母的可行性,并证明了加权后的特征向量具有更高的识别率。
The wavelet decomposition algorithm is applied to extract the structural features of Chinese characters. Chinese charaders are classified according to the main structure. The fractal dimension of the Chinese charaders are rec ognized as the feature vector for recognition. The existing data is used to test the recognition feasibility of English alphabet on a small scale using fractal dimension. It is demonstrated that the weighted feature vector has a better recognition rate.
出处
《新乡学院学报》
2014年第2期21-24,共4页
Journal of Xinxiang University
基金
河南省高等学校人文社科研究项目(2012-qn-219)
河南省科技厅软科学研究项目(132400410966)
河南省政府决策研究招标项目(2013B345)
关键词
文字识别
分形维
小波
多分辨分析
加权
character identification
fractal dimension
wavelet
multi resolution analysis
weight
