3D Brain Segmentation Using Active Contour with Multi Labeling Method

The main objective task in this paper is 3D multi region segmentation based on active contour with region label prior. The region label prior in our algorithm based on measuring the geometrical different in the brain volume. We will scope on level set with grow-cut method for multi regions segmentation regardless to number of regions in the loaded data. The combining of energy minimization method with grow-cut method is very useful in medical image segmentation to extract the desired object regardless to its grey level. Our algorithm use DICOM medical data and it implement in MATLAB 2008 with the aid of 3D slide viewer to visualize 3D segmentation results. The proposed algorithm shows improvements to segment different region of the brain. Our results compared with the results of other papers that considering the brain segmentation and we found the benefit of our proposed algorithm due to its flexibility to select any object in 2D and 3D DICOM medical images.