Medical Microscopic Image Segmentation Study

Immunohistochemistry is an important method displaying chemical ingredients in tissues through particular antibody and it displays experimental results using micrograghs.Computer-assisted image processing analysis has been the trend of immunohistochemical micrograph analysis. With the help of computer image analysis system, we used the method of average positive stained area percentage APSAP to evaluate the slice immunohistochemistry result.It is very important in assistant diagnosis.In this paper,we made the immunohistochemical micrograph of gastric adenocarcinoma with Ki67 expressing as a major object. The extraction and quantificational analsis of positive objects are focused on.In this paper, we describe the basic Image-Segmentation techniques for the medical images,and then discuss and analyze the latest algorithms for the extraction of positive objects.The pathology experts give the diagnosis all by the knowledge base of gastric adenocarcinoma cell characteristics. In order to extract and quantificational analyze accurately, the technology of pre-processing for the gastric adenocarcinoma image , the segmentation of positive objects and the method of average positive stained area percentage have been researched.We put up two algorithms on account of the particularity to this problem.The relationship between the information of the color and the content of the data have been considerd in both of the two algorithms.The method of the segmentation based on YIQ: we converted the primary picture from RGB to YIQ. After the thresholding we marked the interesting area in component I.With the marked, we can extract the objective area. The method of the segmentation based on Euler Distance algorithm: we compute the euler distance between two points, then segment the positive stained area based on color clustering.As is showed in experiments, the results by the new algorithms present in this paper is very good and easily to execute.