Liver CT image segmentation algorithm research based on CV model

Liver CT images are greatly influenced by noise, and the area boundaries of liver image is fuzzy and part of the scanning fragments are scattered and discontinuous, which result in CT image contains some false contour points that affect the process of image segmentation and decrease the accuracy of...

Full description

Saved in:
Bibliographic Details
Published in2017 3rd IEEE International Conference on Computer and Communications (ICCC) pp. 1889 - 1892
Main Authors Shao, Xiangxin, Lin, Xiaomei, Shang, Tingting
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2017
Subjects
Active contour model
Active contours
Computed tomography
CV model
Histograms
Image segmentation
Level set
Liver
Mathematical model
Wasserstein distance
Online AccessGet full text

Cover

More Information
Summary:Liver CT images are greatly influenced by noise, and the area boundaries of liver image is fuzzy and part of the scanning fragments are scattered and discontinuous, which result in CT image contains some false contour points that affect the process of image segmentation and decrease the accuracy of image segmentation. In this paper, a kind of active contour model based on Wasserstein distance is put forward. Wasserstein distance was introduced in the model for comparison in different regions in the image histogram, to enhance the accuracy of similarity measure, calculate the curve near the point of local energy as a model, and the internal energy of the sum so as to improve the accuracy of image segmentation. Experiment shows that the algorithm can be used for accurate segmentation of liver CT images and the energy function has better convergence.
DOI:10.1109/CompComm.2017.8322866