摘要
舌苔腐腻分析是中医客观化、现代化的重要内容 .本文根据腐腻苔的纹理特点 ,提出了一种舌苔腐腻的图像分析算法 .首先将舌苔区分为固定大小的块 ,然后对各块进行分类 ,最后得到整幅图像的腐腻指数 .舌苔块分类时 ,采用改进的子空间法分析纹理结构的疏密 ,并结合纹理粗糙度特征来表达颗粒的粗细 .实验表明其分析结果正确率达到 83% 。
The analysis of curdy and greasy property of tongue fur is of great significance in Traditional Chinese Medicine characterization and modernization.An image analysis method of curdy and greasy fur based on texture feature is proposed in this paper.First,the tongue fur region is divided into blocks with the same size;then the blocks is classified;finally the greasy index of the whole image is calculated.To classify the blocks,a modified subspace method is used to analysed the denseness of the coating texture,combined with the texture coarseness to indicate the fur grain size.Experimental results show that the correctness rate exceeds 83%,and the variances of greasy degree is also reflected.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2003年第z1期2083-2086,共4页
Acta Electronica Sinica
基金
国家自然科学基金 (No .60 2 2 71 0 1
No .60 30 1 0 0 3)
关键词
舌象分析
子空间法
纹理粗糙度
tongue image analysis
subspace method
texture coarseness
