Conference Proceeding talk: Towards Developing a Lightweight Neural Network for Liver CT Segmentation
Date:
This talk focused on the development of Res-PAC-UNet, a cutting-edge neural network tailored for liver CT image segmentation, specifically targeting challenges in the diagnosis and treatment planning of liver diseases like hepatocellular carcinoma (HCC). The network employed Pyramid Atrous Convolutions and a fixed-width residual UNet backbone to achieve high segmentation accuracy with a remarkably low parameter count. This efficient and lightweight architecture demonstrated the potential to enhance clinical workflows by providing precise and scalable solutions for liver CT segmentation, paving the way for improved patient outcomes in liver disease management.