The Application Of Mathematical Morphology In Liver Cancer Cell Pathological Recognition

At present, liver cancer is one of the diseases that afflict human health. According to statistics, the incidence and mortality of liver cancer in recent years began to develop to the younger crowd. Therefore, the prevention and early diagnosis and treatment of liver cancer is the important part that the medical profession pays attention to and the use of technology of computer recognition has become an auxiliary diagnostic method for this part.This paper is mainly based on mathematical morphology( that is morphology) image processing technique to study the pathological slices of liver cells. Take advantage of the characteristics of superior category that the support vector machine can be on two kinds of samples and then identify the images for pathologic slices of liver cells. Due to noise in images for pathologic picture of hepatocytes. It’s found in the process of feature extraction that directly using Canny Edge Detection Algorithm cannot get a closed area, so morphological features cannot be used to perform data analysis for pathologic slices of liver cells. Therefore, based on a single pathological picture of liver cells of morphology of the closed area, we obtain and put forward two methods. One is the combination of the Canny Edge Detection Algorithm and the between-cluster variance method(OTSU).According to the detection effect of the gray scale of the image of pathologic slices, we choose the better one. The other one is that using Three- channel Information Fusion Technique of color image to preprocess the images for pathologic slices of liver cells, mainly on information of color components on R, G, B. Use Canny Edge Detection Algorithm, and then integrate information to get the closed area.Apply morphological edge detection and extract digital information of the closed edge on liver cells in pathological sections, according to extractive boundary, calculate the perimeter, area, shape factor, waviness, and eccentricity, and other features of the cell area. After a lot of experimental analysis, select the shape factor, waviness, and the elongation as the training data of the support vector machine( SVM), Preliminary judge the probability that pathological slices is liver cancer. Assist doctors to diagnose pathology and improve working efficiency.