基于自适应阈值的活体细胞图像分割改进方法

摘要

针对非活体细胞去除问题，提出一种改进的活体细胞图像分割方法。该方法利用自适应阈值分割算法对原图像进行阈值分割，然后对图像中所有连通区域面积特征值进行判断，去除非活体细胞。由于图像初次分割后细胞边界存在不完全封闭的情况，提出了改进的图像边缘断点连接算法和改进的孔洞填充算法，使细胞边缘闭合，从而得到完整的细胞表面图像。实验结果表明，该方法能够准确的进行细胞分割，具有较好的鲁棒性。%An improved method of images segmentation to remove non-living cells is presented in this paper. The method is described as follows:an original cell image is segmented by the adaptive thresholds segmentation, and the area characteristic value of connected region is judged with the purpose of removing non-living cells. When living cells edges are disconnected after segmentation, an improved edge breakpoint connection algorithm and holes filling algorithm are given in order to make the closed cell edge and obtain the whole cell surface images. The experimental results show that the method has good robustness and accuracy.