Retinal vasculature image segmentation algorithm based on deep learning

Nanxi Jin; Yuelan Xin; Shenghui Deng; Kuan Chen; Lan Yang; Haoyi Huo; Jia Zhao; Huiting Fang

doi:10.1117/12.3040462

26 September 2024 Retinal vasculature image segmentation algorithm based on deep learning

Nanxi Jin, Yuelan Xin, Shenghui Deng, Kuan Chen, Lan Yang, Haoyi Huo, Jia Zhao, Huiting Fang

Proceedings Volume 13282, Second Advanced Imaging and Information Processing Conference (AIIP 2024); 1328204 (2024) https://doi.org/10.1117/12.3040462
Event: Second Advanced Imaging and Information Processing Conference (AIIP 2024), 2024, Xining, China

Abstract

To enhance the accuracy of existing algorithms in the task of retinal vessel image segmentation, this paper proposes the incorporation of two modified convolutional blocks, in lieu of traditional ones, within the framework of the U-Net neural network, aiming to strengthen the extraction of detailed features. Firstly, a convolutional block equipped with local channel attention is devised for feature extraction in the shallow layers of the network. Secondly, a convolutional block incorporating global channel attention is introduced for feature extraction in the deeper layers. Lastly, skip connections are employed to feed the features extracted from the shallow layers into the deeper layers. Experimental results demonstrate that the proposed retinal vessel segmentation algorithm achieves a Dice coefficient of 88.21% and a sensitivity of 87.16%, marking improvements of 2.25% and 1.64% respectively over the original U-Net network.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Nanxi Jin, Yuelan Xin, Shenghui Deng, Kuan Chen, Lan Yang, Haoyi Huo, Jia Zhao, and Huiting Fang "Retinal vasculature image segmentation algorithm based on deep learning", Proc. SPIE 13282, Second Advanced Imaging and Information Processing Conference (AIIP 2024), 1328204 (26 September 2024); https://doi.org/10.1117/12.3040462

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available