摘要
舌色的模式识别是为了促进舌诊医学的现代化。舌图一般用RGB或HSV这两种颜色特征来区分。采用主成分分析(principal components analysis,PCA)对颜色特征进行处理,获得两种新的特征,即RGB-PCA和HSV-PCA。再用支持向量机(support vector machines,SVM)和误差反向传播神经网络(baek propagation neural network,BP NN),分别对四组特征值进行识别研究,找出合适的算法参数,在此基础上比较了支持向量机和BP神经网络的识别结果。
Pattern recognition is applied to tongue images to improve glossoscopy researches. The color features of tongue images can be characterized by color spaces of RGB or HSV system or projected PCA space, using Support Vector Machines and Back Propagation neural network methods. Parameters of these methods are also investigated in order to obtain better results.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2008年第6期721-723,共3页
Computers and Applied Chemistry
关键词
舌诊
支持向量机
神经网络
主成分分析
颜色
glossoscopy, support vector machine, back propagation neural network, principal components analysis, color
