Paper Info
Title
An improved multilevel banded graph cuts method for extracting colon tissue
Abstract
This paper proposes an improved multilevel banded graph cuts method for colon segmentation. Accurate and quick colon extraction from 3D medical abdomen CT images is a crucial step in virtual colonoscopy. However, it is a challenge to extract colon tissues because of the local characteristics of colon tissues and multi-megapixel 3D medical images. Multilevel banded graph cuts method has drastically reduced the running time, but the segmentation accuracy needs to be improved. This paper utilizes Gaussian pyramid technologies for the high-resolution image down-sampling. The loss of image information can be reduced and the noises can be treated effectively. The graph cuts method is performed on banded graph directly and the terminal seed points are adjusted dynamically. Meanwhile, different weighting functions are used to describe the similarity degree among pixels. Partial volume effects are eliminated by a closing operation in morphology. Finally, this paper implements the three-dimensional visualization of the final segmented colon tissues and compares the experimental result with the standard data. The segmentation accuracy of our proposed method can reach 95.9%.
Year
DOI
Venue
2015
10.1109/FSKD.2015.7381970
2015 12th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD)
Keywords
Field
DocType
colon segmentation,graph cuts,banded graph,weighting function,partial volume effects
Cut,Weighting,Scale-space segmentation,Computer science,Segmentation-based object categorization,Image segmentation,Artificial intelligence,Virtual colonoscopy,Computer vision,Pattern recognition,Segmentation,Pixel,Machine learning
Conference
Citations 
PageRank 
References 
0
0.34
9
Authors
9
Name
Order
Citations
PageRank
Wei He103.04
Liyuan Zhang200.34
Huamin Yang31917.29
Zhengang Jiang4226.42
WeiLi Shi515.10
yu miao647.18
Fei He700.34
Fei Yan811115.03
Huimao Zhang901.01