【摘要】 针对人耳识别中因人耳与肤色部分颜色接近,对比不鲜明而导致的很难进行边缘检测的问题,该文提出了一种新的外耳轮廓跟踪算法。算法首先运用阈值分割和形状匹配的方法找到外耳轮廓顶点,并将外耳轮廓分为左右两部分。从顶点开始,通过四个方向的Prewitt算子和一个最佳方向模板首先确定外耳轮廓的基本走向,然后分别从左右两个方向进行轮廓跟踪,得到外耳轮廓。将该算法与其它算法进行比较实验,结果表明了该算法的有效性。更多还原

【Abstract】 It’ s hard to detect outer ear contour,as the color of ear is similar to that of face. In order to settle the problem,a novel algorithm of outer ear contour tracking is proposed in this paper. First,get the top position of the outer ear contour by threshold seg-mentation and shape matching. Second,determine its direction by four Prewitt templates and one best template. Then,the outer ear contour is divided into two sides,left side and right side. Detect the outer ear contour from the top position on both sides. Finally,compare the proposed algorithm with others. The result demonstrates the validity of the algorithm.更多还原