Image Segmentation and Classification for Medical Image Processing

Abstract

Segmentation and labeling remains the weakest step in many medical vision applications. This paper illustrates an approach based on watershed transform which are designed to solve typical problems encountered in various applications, and which are controllable through adaptation of their parameters. Two of these modules are presented: the lung cancer detection, a method for the segmentation of cancer regions from CT images, a watershed algorithm for image segmentation and brain tumor detection from MRI images. Various GLCM features along with some statistical features are used for classification using Neural network and Support Vector Machine (SVM). We describe the principles of the algorithms and illustrate their generic properties by discussing the results of both applications in 2D MRI images of Brain tumor and CT images of lung cancer.