Abstract: Analyzing the medical image in image processing is the most important research area. Capturing the image are analyzed to identify different medical imaging problems is the common factor in this field. Robust organ segmentation is a prerequisite for computer-aided diagnosis (CAD), quantitative imaging analysis, pathology detection and surgical assistance. Some of the organs in the human body have high anatomical variability, so segmentation of such organs is very complex. The proposed system segments the pancreas with the considerations of spatial relationships of splenic, portal and superior mesenteric veins with the pancreas. The proposed system uses macro super-pixels for fast and deep labeling and segmentation process. The proposed system is an automated bottom-up approach for pancreas segmentation with the consideration of spatial relationships with the veins in abdominal computed tomography (CT) scans. The method generates dynamic cascaded and macro super-pixel segmentation informationís by classifying image patches at different resolutions. Fast organ analysis using Dense-SIFT algorithm.

Keywords: Image processing, medical image, pancreas segmentation, CT and Dense-SIFT algorithm.