Research, Medical Image Processing Based On Mathematical Morphology

The microscope used to analyze immunohistochemical result may play an important role, but subjective analysis can affect the accuracy of the results. With the development of computer technology, computer image processing and analysis techniques play an increasingly important role in clinical diagnosis and treatment. The quantitative analysis of cell images by computer used to assist the doctor to make faster and accurate judgments is more important in the future. This paper focuses on the extraction of positive and negative target areas and the quantitative analysis.First of all, this thesis introduces an overview of the basic theory of morphological image processing and enhances the contrast of images with the technology of Morphological image enhancement in color microgram. Then, it carries out color space conversion of the images and analyzes the color of the images with the method of Colorimetric Criterion. The research shows that in positive cell images, each R-component image pixel is larger than B-component; in negative cell images, each R-component image pixel is smaller than B-component. Then, the colorimetric criterion crudely segments immunohistochemical cell image, and we can get the images that contain positive cell and negative cell. Moreover, the medical image segmentation was proposed roughly, including a variety of segmentation methods. In this thesis, it applies the C-means clustering segmentation algorithm to crudely divide the positive cell images and negative cell images. Finally, after the Morphological filter processing to positive cell images and negative cell images, we abstract the feature information such as the number and area of cell images.The experimental results show that C-means clustering segmentation algorithm has a good effect on the division of the immunohistochemical images.