An Improved Segmentation Method for Color Image

In the color image, based on the characteristics of the HIS (Hue-Saturation-Intensity) color space, there is most of the white blood cell information in the H component of the HIS color space, and there is most of the white blood cell nucleus's structure information in the S component, we develop an improved method, iterative GGM (Gray level and Gradient Mapping) function segmentation based on circular histogram for the leukocyte segmentation by making full use of this knowledge in stead of the iterative Otsu's approach based on circular histogram for the leukocyte segmentation. We call it iterative GGM function based on circular histogram method for short. The result of simulation demonstrates that, the new algorithm described in this paper can also detect the cell and cell nuclei effectively. The algorithm may have a widespread application prospect.