One Kind of Macrophages Images Segmentation and Labeling Method

Segmentation and labeling are the important preprocessing steps towards pattern recognition and image understanding and significant step towards image compression and coding. To overcome the difficulty caused by the overlapping of macrophages in the image automatic analysis, the touching macrophages need to be separated into single ones. This paper presents a novel method of macrophages image segmentation and labeling. The image first converted colors form RGB to YIQ. Then the erosion and dilation of morphological filter processing are used to get rid of noises. And then the chamfer distance transformation converts binary digital image into gray-level image. The watershed algorithm is carried out to segment the macrophages image. Finally, we use the sequential labeling algorithm which scans the image twice to label macrophages. Experimental results show this method works successfully in the segmentation and labeling of macrophages microscopic images which are get from blue channel.