Can we make a more efficient U-Net for blood vessel segmentation?

Qiong Liu; Ziming Zhong; Sourya Sengupta; Vasudevan Lakshminarayanan

doi:10.1117/12.2567861

20 August 2020 Can we make a more efficient U-Net for blood vessel segmentation?

Qiong Liu, Ziming Zhong, Sourya Sengupta, Vasudevan Lakshminarayanan

Proceedings Volume 11511, Applications of Machine Learning 2020; 115110I (2020) https://doi.org/10.1117/12.2567861
Event: SPIE Optical Engineering + Applications, 2020, Online Only

Abstract

Blood vessel segmentation is an important step in the automated diagnosis of ophthalmic disease from retinal fundus images. The UNet is a popular encoder-decoder architecture widely used in biomedical pixel-wise seg- mentation problems. In this paper, we analyze how the UNet can be used in a more computationally efficient way. Pre-trained weights are used to initialize the network and 3 different architectures are used to compare and analyze the efficacy of the models in terms of both computational cost and performance. Three different deep architectures (VGG16, ResNet34, DenseNet121) are discussed and their efficiencies are compared for the blood vessel segmentation task. Resnet34 architecture achieved highest sensitivity of 0.849 and accuracy and specificity of 0.961, 0.9843 with number of parameters as low as 510178 compared to normal UNet with 34525168 parameters and a sensitivity of 0.756.

Conference Presentation

Citation Download Citation

Qiong Liu, Ziming Zhong, Sourya Sengupta, and Vasudevan Lakshminarayanan "Can we make a more efficient U-Net for blood vessel segmentation?", Proc. SPIE 11511, Applications of Machine Learning 2020, 115110I (20 August 2020); https://doi.org/10.1117/12.2567861

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