基于Grab Cut和八方向链码法的藻类细胞轮廓提取算法*

作者：陈 浩,庞 全 日期：2010-09-08/span> 浏览：3783 查看PDF文档

基于Grab Cut和八方向链码法的藻类细胞轮廓提取算法*

陈 浩,庞 全

（杭州电子科技大学 生物医学工程与仪器研究所,浙江 杭州 310018）
摘要：为了解决藻类细胞轮廓提取的困难，提出了一种不同于传统方法的新的边缘提取算法。首先对图像进行Grab Cut处理，然后进行阈值法二值化，接着用八方向链码法轮廓跟踪，最后用傅立叶描述子进行边界平滑。实验结果表明，该算法不仅忽略了藻类细胞内部杂乱的纹理，还较好地保留了细胞边缘信息，精确地提取了外部轮廓。
关键词：Grab Cut；八方向链码法；傅立叶描述子；藻类图像分割
中图分类号：TP391.41
文献标识码：A文章编号：1001-4551(2010)08-0108-04

Algae cell contour extraction algorithm based on Grab Cut and the 8-direction chain code edge tracking method
CHEN Hao, PANG Quan
(Institute of Biomedical Engineering and Instrument, Hang Zhou Dianzi University, Hangzhou 310018, China)
Abstract: To solve the problem in algae cell contour extraction, a novel algorithm of image edge detection which is different from the existing algorithms was proposed. Firstly, the Grab Cut algorithm was applied to the filtered image. Afterward, the Threshold method was used to get a binary image. Then, the 8-direction chain code edge tracking algorithm was used to track the edge，at last Fourier descriptors were applied for boundary smooth. Experimental results show that the algorithm can not only neglect the complex inner texture but also preserve the edge information of algae cell completely, precisely extract the external contour.
Key words: Grab Cut; 8-direction chain code edge tracking algorithm; Fourier descriptors; algae cells image segmentation