AUTHOR=Lu Haoran , She Yifei , Tie Jun , Xu Shengzhou 

TITLE=Half-UNet: A Simplified U-Net Architecture for Medical Image Segmentation

JOURNAL=Frontiers in Neuroinformatics

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroinformatics/articles/10.3389/fninf.2022.911679

DOI=10.3389/fninf.2022.911679

ISSN=1662-5196

ABSTRACT=Medical image segmentation plays a vital role in computer-aided diagnosis procedures. Recently, U-Net is widely used in medical image segmentation. UNet3+ further explores sufficient information from full-scale features, which not only improves accuracy but also reduces network parameters. In this paper, the effects of different parts of U-Net on the segmentation ability are experimentally analyzed. Then a more efficient architecture named Half-UNet is proposed. The proposed architecture is essentially an encoder-decoder network based on U-Net, in which both encoder and decoder are simplified. The re-designed architecture takes advantage of unification of channel numbers, full-scale feature fusion and Ghost modules. We have evaluated Half-UNet with U-Net and UNet3+ architectures across multiple medical image segmentation tasks: mammography segmentation, lung nodule segmentation in the CT images, and left ventricular MRI image segmentation. Experiments demonstrate that, compared with U-Net and UNet3+, Half-UNet has similar segmentation accuracy, while with at least 98.6% and 98.4% fewer parameters, 81.8% and 95.3% fewer FLOPs.