摘要： 多细胞粘连识别是图像识别领域的瓶颈问题，为此，提出一种多细胞粘连的形状识别方法。通过改进的阈值分割和八方向边界描述法对细胞进行定位，基于多边形夹角法分离各个粘连细胞，解决了多细胞粘连的形状识别问题。实验结果表明，该方法适用于多细胞复杂粘连情况下的细胞分离和各个细胞的形状识别，系统粘连分离精度为95.3%，形状识别精度达到97.6%，每个细胞识别时间为0.03 s。

关键词: 多细胞粘连, 形状识别, 改进的阈值分割, 八方向边界描述, 多边形夹角法

Abstract: Multi-cell adhesion shape recognition is a key problem in the field of image recognition. The improved threshold segmentation and eight direction boundary description strategies are proposed to locate the cells in this paper, and then the shape of those cells is recognized by separating each adhesion based on polygonal angle-offset method. Experimental results show that this method is appropriate to the complex situation where multi-cell conglutination separation and each cell’s shape recognition are needed. This system’s multi-cell conglutination separation accuracy is 95.3%, shape recognition accuracy is 97.6%, and each recognition of cell costs 0.03 s.

Key words: multi-cell adhesion, shape recognition, improved threshold segmentation, eight direction boundary description, polygonal angle- offset method

中图分类号: 

• TP391.4