基于改进的最大类间方差的舌色和舌苔分离技术

舌苔、舌色分离是实现舌诊图像自动分析的重要的预处理工作。舌的分部图象的直方图是单峰的，采用一维最大类间方差法确定最佳分割阈值比较合适的，该方法简单有效，很受欢迎。但由于其对噪声和图象频数分布的敏感性，成为当前研究的热点之一。该文提出一种新计算公式，实验表明该公式对噪声等零星分布有较强的抑制作用，得到的最佳阈值比较得稳定，有参考意义。

Based on Advenced Otsu Methord Separating the Tongue Coating and the Tongue Image

In automatic analysis of the tongue n-nage, to separanng the tongue coanng ana me tongue Lrnage is a n-nportant pretreatrnent work.The OTSU method is a simple and effective method for image segmentation, especial when the histogram of which image has single peak value.So it is very useful,to process the tongue image by OTSU method.But which is semitivity to the noise and image frequency dis- tribution, And becomes into a hot spot now.A new formula is put forward in this paper, and the experiment indicated that the fragmentary distribution ＂noise＂ in histogram are inhibited strongly by the new method.The best threshold value get by this paper become more steady, and the method has the reference significance.