Gastric Adenocarcinoma Microscopic Image Segmentation Algorithm Research

With the wide application of computer, it's an important topic in the field of medicine image to make research on Cancer Cell image recognition with Computer Patter Recognition technique. And this paper focuses on the computer aided-diagnosis system for gastric adenocarcinoma with the microscopic cell images which are stained by HE.By summing the experiences of the previous jobs, we choose six parameters of veins, three color character parameters and six morphological parameters. For constructing the aided-diagnosis system, these parameters will be used to decide the cells which are cancers. However, there are so many types of cancer cells and these images are complex. Therefore, it is not easy to fulfill image recognition with this technique. On the base of studying the classical methods, the paper proposes some new ideas.First, for segmenting the nuclei of every cell and extract the features, the paper proposes an FCM algorithm bases on the improved degree of membership. The conventional FCM algorithm is noise sensitive because of not taking into account the spatial information, and the novel modified FCM algorithm is formulated by incorporating the spatial neighborhood into the standard FCM algorithm. A prior probability is given to indicate the spatial influence of the neighboring pixels, and the probability is automatically determined in the implementation of the algorithm by the fuzzy memberships of the neighboring pixels. The new method is applied to real images and is shown to be effective and more robust to noise than the conventional FCM algorithm.Second, in order to separate the images of clustered cells, an improved algorithm based on max separation zones is proposed. For resolving the problems of the iterative erosion, the paper uses different method from conventional morphology to get the new kind of structuring elements; with the right seed points and the max separation zones, the algorithm will get the accurate contours of every cell.These new algorithms improve the conventional methods' shortcomings. Experimental results reveal that improved FCM algorithm is able to get the accurate separation cell, and the reasonable contour is gained by the max separation zones algorithm.Finally, we reconstruct the Database of the aided-system, and the Database will have lots of data to teach students and support the in-depth research.