一种基于Snakes模型的中医舌象分割改进算法  

An Improved Snakes Model for Tongue Image Segmentation

作　　者：王明英[1] 张新峰[1] 卓力[1] 

机构地区：[1]北京工业大学信号与信息处理研究室,北京100124

出　　处：《测控技术》2011年第5期32-35,共4页Measurement & Control Technology

基　　金：北京市自然科学基金资助项目(4092009)

摘　　要：舌象分割是中医舌诊客观化研究中的一个重要步骤,其结果好坏直接影响到舌象自动分析结果。传统的Snakes舌体分割算法的颜色依赖性太大,当舌象的采集环境发生变化时,该算法对所采集的舌象分割结果往往不能令人满意。本文提出的基于Snakes模型的改进算法将舌象转换到多种颜色空间,两次应用Snakes动态轮廓模型,最大化地提取了舌体的颜色特征,在研制的新型舌象仪中,舌象分割取得了较好的分割结果。As tongue image segmentation is an important procedure in the tongue characterization,its accuracy affects the following automatic process directly.The traditional segmentation methods based on Snakes model rely on the color features of the tongue image seriously.When the acquisition environment has changed,the results obtained by the original methods are not usually satisfied.An improved approach based on Snakes model is adopted,which describes the tongue image in two other color space,uses a two-step Snakes implementation from coarse to fine,and collects the color features maximumly.At last,a satisfied result is obtained in the new environment.

关 键 词：舌体分割 SNAKES模型 颜色空间 大津(Ostu)法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]