基于分水岭变换的彩色细胞图像分割

提出了一种彩色细胞图像的分割方法,重点解决邻接或者重叠的细胞核或者细胞质之间的分割问题。算法主要基于数学形态学方法,首先利用k-均值聚类和直方图上的分水岭变换对像素进行聚类,从而得到细胞图像背景、细胞质和细胞核的粗略估计;然后利用形态学滤波和分水岭变换提取细胞核、细胞质的边界。结果表明:在少量人工干预的情况下,所述算法获得的细胞核、细胞质边界与实际情况比较吻合。对于含有邻接或者重叠的细胞核或者细胞质的图像,算法仍能提取理想的细胞组织边界。

Color cellular image segmentation scheme based on watershed transformation

This paper presents an efficient scheme for color cellular microscopic image segmentation.The scheme was developed for segmenting neighboring or overlapping nuclei or cytoplasm.The scheme is based on mathematical morphological methods with coarse pixel clustering achieved using the k-means clustering method and the watershed transformation of the image histogram.Morphological filtering and the watershed transformation were employed to accurately define edges of the nucleus and cytoplasm.Experimental results demonstrate the effectiveness of the scheme for extracting precise,continuous edges with limited human intervention,especially for images with neighboring or overlapping nuclei or cytoplasm.