High effective medical image segmentation with model adjustable method

An integrated algorithm framework for high effective medical image segmentation is proposed in this paper. The proposed framework consisting of four optimal and logic correlative calculation modules is model-adjustable according to a specific segmentation target. Two major applications of the integrated algorithm framework are explored for effectiveness verification. The boundary-based active contour model with priori shape constraint is very suitable for segmenting regions of interest (ROI) of the image with low contrast, blur or occlusion. The hierarchical M-S model combined with diffusion filter is mainly employed for multi-object segmentation of the noisy image. A target-adaptive scheme is preliminarily designed for adjusting the model to resolving a particular image processing task. Experimental results show excellent effectiveness for MR brain image segmentation under different conditions of image quality degradation.