High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

Automatic Fast Pancreas Segmentation and Classification of Abdominal CT Scans

Author(s):

V. Santhoshkumar , Anna University, CSE Department, KIT- Kalaignarkarunanidhi Institute of Technology, Coimbatore, Tamilnadu. ; Dr. R. Adaline Suji, Anna University, CSE Department, KIT- Kalaignarkarunanidhi Institute of Technology, Coimbatore, Tamilnadu.

Keywords:

Computed Tomography (CT) Scans, Segmentation

Abstract

Medical image segmentation is the most popular research area nowadays. In the existing system, the pancreas segmentation has been performed using the superpixel concepts. The proposed system segments the pancreas from the abdomen CT scan images with the new set of algorithms. This system uses the macro super-pixels for fast and deep labeling, segmentation process and proposes an automated bottom-up approach for pancreas segmentation in abdominal computed tomography (CT) scans. The earlier work on organ segmentation achieved only low accuracies when comparing to organs like the heart, liver, pancreas, etc. In this proposed system a complete self-learning deep analysis method is presented for pancreas segmentation. This utilizes the abdominal computed tomography (CT) scans as input. The method developed a segmentation technique by sorting image patches at different resolutions and cascading macro super-pixels. Macro superpixel segmentation, grouping both strength and likelihood features to form observed statistics in pouring random forest frameworks. At last effortless connectivity based post-processing were done. The proposed system utilizes some histogram and texture features. The method generates dynamic cascaded and macro super-pixel segmentation. For Fast organ detection SIFT (Scalable Invariant Feature Transform) algorithm is proposed. This algorithm helps to detect the descriptors from the image for fast image segmentation. This system used some CT scan images and developed segmentation algorithm using Matlab. The organ from the CT images can be segmented quickly and finally the classification task is performed over the segmented pancreas to categorize it. The classification task finds the organ into two classes such as affected or healthy pancreas.

Other Details

Paper ID: IJSRDV6I21557
Published in: Volume : 6, Issue : 2
Publication Date: 01/05/2018
Page(s): 3660-3666

Article Preview

Download Article