基于形状先验水平集的高鲁棒性中医舌图像分割方法

High Robust Tongue Image Segmentation Algorithm Based on Active Contour Model With Shape Priors

摘要: 中医舌图像分割一直面临着由于采集设备更新换代等因素带来的分割方法鲁棒性下降的问题.提出一种基于形状先验主动轮廓模型的中医舌图像分割方法.首先, 采用统计方法获得大致舌体图像;然后, 采用形状先验水平集方法得到基本完整的舌体;最后, 采用Snake模型方法得到最终准确分割结果.实验结果表明:该方法对于不同环境下得到的舌图像分割都具有良好的分割效果.

Abstract: There always exits a problem with the tongue image segmentation that the robustness of the method decreases as the hardware is replaced or upgraded.In this paper, a novel tongue image segmentation method is proposed based on active contour model with shape priors.The statistical method is used to get a rough result, then the level set method with shape priors is used to get a mostly complete tongue, and the exact tongue is finally obtained using a post-processing step.Experimental results show that accurate segmentation can be obtained with the proposed method for different tongue image acquisition environments.