A Novel Lesion Segmentation Algorithm based on U-Net Network for Tuberculosis CT Image

Wen, Shaoyue and Liu, Jing and Xu, Wenge (2021) A Novel Lesion Segmentation Algorithm based on U-Net Network for Tuberculosis CT Image. In: 2021 International Conference on Control, Automation and Information Sciences, 14th - 17th October 2021, Xi'an, China.

[img]
Preview
Text
21_ICCAIS_LungCT.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (1MB)

Abstract

Lung CT images provide several essential information for lung disease diagnosis and lung surgery. However, the traditional detection method through manual segmentation is laborious and time-consuming. This paper presents automatic tuberculosis (TB) lesion segmentation method based on U-Net neural network for detecting TB. In addition, we combined an edge detection algorithm called canny edge detector with this network to get a more accurate TB lesion boundary. This method is trained on two split databases with 3576 lung CT images obtained by data enhancement on 447 discontinuous lung CT images. The results show that the proposed approach is validated for complex TB lesions with a high dice coefficient (91.2%).

Item Type: Conference or Workshop Item (Paper)
Identification Number: https://doi.org/10.1109/ICCAIS52680.2021.9624633
Dates:
DateEvent
30 August 2021Accepted
9 December 2021Published Online
Uncontrolled Keywords: Image processing, lesions segmentation, Unet, tuberculosis
Subjects: CAH02 - subjects allied to medicine > CAH02-05 - medical sciences > CAH02-05-01 - medical technology
CAH11 - computing > CAH11-01 - computing > CAH11-01-01 - computer science
CAH11 - computing > CAH11-01 - computing > CAH11-01-02 - information technology
Divisions: Faculty of Computing, Engineering and the Built Environment > School of Computing and Digital Technology
Depositing User: Wenge Xu
Date Deposited: 05 Oct 2021 09:58
Last Modified: 22 Mar 2023 12:00
URI: https://www.open-access.bcu.ac.uk/id/eprint/12253

Actions (login required)

View Item View Item

Research

In this section...