| The cell classification and recognition are an important research issue in imaging medicine. Segmentation of a cell image is to distinguish an cell image into different pats according to distinct properties and meaning of sub-regions of the cell image, it is a key step in cell analysis and recognition. After segmentation the sub-regions of the cell image are separated and uncrossed each other, possessing the certain properties. The cell kinds and cell pattern of human being bodies are numerous, and quantities of different cell images are of obvious difference. The requires for segmentation quality of cell image are very high and serious. So, the segmentation of cell image is an important and difficult task in cell image processing.A color cell image possesses more abundant information than a black-white image or gray cell image. Many classical segmentation algorithms of cell image are based on binary image or gray image. The research of cell image segmentation based on the color cell image is becoming an active research area.Based on the characteristics of dying cell image the paper proposes a segmentation method of color cell image. Choosing the HSI color space for suppressing the background noises, combining the auto-threshold and color hue degree for removing background, the method mainly applies the fast water immersion algorithm for image segmentation of a color cell image. The middle-value filter and homogeneous processing were used for overcoming the over-segmentation, and region merge algorithm was improved according to features of the color cell image. The segmentation results of cell image were correct and accurate. The definition and developing situation of image segmentation were briefly introduced, some main segmentation algorithms were briefly described first. Then,the color image space and the main idea and computing method of the fast water immersion algorithm were described. Finally, the computing flow procedure and experiment results were provided and discussed.
|
PDF Full Text Request