基于ECCC的细胞图像分割算法

ECCC algorithm on image sequences of stem cells

针对粘连细胞图像,提出ECCC(Eleven Components Chain Code)链码分割算法。首先对细胞边缘二值图像进行链码统计和边缘拐点检测,新算法对Freeman链码进行了改进,在链码中加入表示边缘拐点的新的链码元素,然后计算边缘拐点的链码差来筛选真实分割点,最后对分割点线性插值实现粘连细胞分割。实验结果表明,针对2粘连和3粘连细胞,ECCC法的分割成功率分别为100%和98%,平均耗时分别为0.42 s和0.67 s,比传统链码分割法减少了近55%的计算量,在复杂的细胞图像分割中具备一定的有效性和可行性。

For the adherent cells,ECCC（ Eleven components Chain Code） is proposed.Firstly the chain code is counted and the cor-ner is detected based on the pre-treated edge binary image.The improved algorithm solves the problem of high computation com-plexity in the traditional Freeman 8 neighborhood chain code mode by adding a new chain code element that represents the edge inflection point in the chain code.Then the real segmentation points are detected by calculating the chain code difference of the inflection point in the cell edge.Finally,the adherent cells are separated by linear interpolation of real segmentation points.For two adhesion and three adhesion cells,this method is proved by experimental results to have higher computering speed and better de-noising performance than traditional chain code method.The success rate of the segmentation algorithm is 100 % and 98 % respec-tively and the average consumption time is 0.42 s and 0.67 s respectively.ECCC has certain validity and feasibility in the complex cell image segmentation.