基于高光谱成像技术的中医舌纹分析算法

针对中医舌诊现代化中使用彩色或灰度图像进行舌纹分析困难的问题,采用高光谱成像系统代替数码相机进行舌图像采集,并提出基于Gabor滤波器的舌纹分析算法.这一算法充分利用高光谱舌图像图、谱两方面的信息进行舌纹分析,并根据特征向量之间归一化的距离对部分典型舌纹进行初步分类.在对474例具有典型舌纹的高光谱舌图像的分类中,除了在来蛇纹与去蛇纹、四直纹与曲虫纹、太阳纹和龟纹、左右撇纹和锯齿纹之间有一定的误分外,整体具有较好的分类精度.这些分类结果表明,基于高光谱图像的舌纹分析方法明显优于基于灰度图像的舌纹分析方法.

Tongue texture analysis and discrimination algorithm based on the hyperspectral image

Tongue diagnosis is an important diagnostic method in traditional Chinese medicine (TCM). However, due to its qualitative, subjective and experience-based nature, traditional tongue diagnosis limits application in clinical medicine. To solve the problems of current tongue texture analysis methods, a hyperspectral imaging system instead of a digital camera was used to capture tongue images and a novel algorithm for automatic tongue texture analysis based on the Gabor filters was presented. In this algorithm, tongue texture features were derived from both individual bands and combinations of bands of the hyperspectral tongue images. The automatic tongue texture analysis and discrimination results of the new method are better than the results based on the digital camera images.