赤足跖后缘自动提取与自适应变模型特征识别

将经典Snake模型与自适应变模型相结合的算法，快速提取图像中的开边缘并识别边缘形状，实现赤足足迹跖后缘的自动提取与形状识别。首先，对原图像校正、平滑去噪，由数学形态学方法得到Snake模型的初始轮廓，迭代求解实现跖后缘的快速、精确提取。然后，由高次多项式进行数字滤波，再对滤波后的边缘曲线，利用先验知识，依据平均曲率的变化，自适应地选取不同的函数模型，提取出相应特征，自动识别边缘形状。该方法边缘提取准确，特征提取合理，使计算机总体识别率达到90％。

Automatic edge detection of barefoot metatarsus region and adaptive changeable models for feature identification

To detect opening edges and carry out shape identification,an approach which combines original Snake algorithm with adaptive changeable models is proposed.The process includes image rectification,noise smoothing,initial outline of Snake model was obtained.Then using polynomial as digital filter,it gains smooth curve function to depict the metatarsus region edge.Finally,using priori knowledge,the method selects different function models adaptively based on average curvature to compute corresponding characters and judge the shapes of edges automatically.Results show that the method is rapid and effective.