Study On Intelligent Method For Color White Blood Cell Image Segmentation

Because of the main idea of image segmentation is to simplify and change imageinto easier and meaningful form to analyze, image segmentation is first and one of themost important steps in image analysis. With the development of computer-assisteddiagnosis, Medical image analysis highlights its importance has become increasingly inthe modern medical pathological analysis. Cell image segmentation can determine thecontour and position of the cell so that target cell is separated, which means cell featureextraction and cell recognition is more accurate and getting a more accurate analysis ofpathological. Modern medical equipment requires intelligence which requires imagesegmentation is intelligent. Smart cell image segmentation automatically and accuratelycomplete the separation of the cells, promoting the intelligence, automation andreal-time of the entire cell image processing system.Exact color image segmentation, on the one hand, depends on the appropriate colorspace, on the other hand is the correct choice of the segmentation method. The choice ofcolor space is the first step of the color image segmentation. Selected the appropriatecolor space, the color image segmentation can be more easily accomplished and bettersegmentation result can be gotten. Simply select the appropriate color model for colorcell image; it can take advantage of simple threshold method for image segmentation,and accurate segmentation results. As for color cell image, as long as the choice of theappropriate color model, image segmentation can simply use threshold segmentationmethod to be accomplished for a precise one. In view of this situation of the color whitecell image segmentation and according to the actual test and analysis of various colorspace, four kinds of color components in the color space (HSI, HSB, LAB, gray) arealternatively selected, and use intelligent method to choose one suitable component tosegment the white blood cells image.In order to obtain intelligent selection for the above four color models components,two algorithms of active judgment based-model and segmentation judgmentbased-model methods are presented and tested contrastively. Active judgment mode ismainly based on the different color model image statistical characteristics to get directdecision algorithm. In this paper, intelligent image segmentation of the white blood cellsis the study of how to intelligently select the appropriate color space components. As forsegmentation method, Threshold is enough, so the mean and variance are naturally usedas image statistical characteristics. Based on segmentation evaluation，segmentation judgment model get indirect decision algorithm. Segmentation judgment model presentstexture consistency judgment segmentation correctness and histogram valley detectionjudgment segmentation is correct or not, respectively. The three kinds of intelligentselection algorithm on the segmentation of white blood cell image intelligence testsshow active judgment based-model can finish color model initiative-selection to getintelligent white blood cell image segmentation in a certain extent and segmentationjudgment based-model can complete segmentation intelligence-judgment easily tosegment white blood cell image intelligently.