The Study On The Tumor Microscopic Image Segmentation Technology Based On Snake Mode

The medical micro-image is an important basis for clinical doctors and experts to diagnose the disease. Medical image segmentation technology that is used to extract the target interested for doctors to clinical diagnosis and medical research. Especially, the comparison of the shape and size changes of cells in micro-image under different pathological states,can provide new scientific foundation for pathological analysis and disease diagnosis. So, the accuracy of image segmentation direct impact doctors to determine the true disease and make the right diagnosis. However, the traditional technology of image segmentation is difficult to achieve the desired results due to the complex organizational structure and shape of medical microscopic image, together with other factors such as ambiguity and uneven, image noise, image quality and so on.In view of the limitations of traditional image segmentation technology, this thesis provided an new image segmentation method based on Snake model. First of all, this thesis introduced the major image segmentation technology, and analysed the segmentation results of commonly medical image segmentation approach on the gastric tumor microscopic image, Secondly, this thesis discussed the dynamic image segmentation technology-Active Contour Models. It is a dynamic curve guided by the internal and external energy that pull in toward the border of objects under the principle of energy minimization. In this section, this thesis described the basic principles, mathematical definition and calculation methods of Active Contour Models. After the research of the image features of this subject, this thesis applied the technique of Active Contour Models based on the gradient vector flow(GVF Snake) to the cell segmentation method. In the experiments, this thesis provided an improved Active Contour Models based on GVF in order to overcome the disadvantages of GVF Snake. It is proved that the higher accuracy and efficiency of this improved GVF Snake by the results of segmentation experiments. Finally, the new segmentation method is applied to the tumor image processing software.