Automatic Liver Segmentation on Volumetric CT Images Using Supervoxel-Based Graph Cuts

• DOI: 10.1155/2016/9093721
• Publication Date: 2016-04-05T17:02:51Z
• Authors: Weiwei Wu; Zhuhuang Zhou; Shuicai Wu; Yanhua Zhang
• Keywords: Cut; Maximum intensity projection

Accurate segmentation of liver from abdominal CT scans is critical for computer-assisted diagnosis and therapy. Despite many years research, automatic remains a challenging task. In this paper, novel method was proposed delineation on volume images using supervoxel-based graph cuts. To extract the interest (VOI), region abdomen firstly determined based maximum intensity projection (MIP) thresholding methods. Then, patient-specific VOI extracted by histogram-based adaptive morphological operations. The supervoxels were generated simple linear iterative clustering (SLIC) method. foreground/background seeds cuts largest slice, algorithm applied to supervoxels. Thirty used evaluate accuracy efficiency algorithm. Experimental results show that can detect accurately with significant reduction processing time, especially when dealing diseased cases.