Final project for TJ-100685 深度学习 | Deep learning, Sep. 2019 - Jan. 2020
Utilized the latest Weight Standardization}** (WS) as well as GroupNorm to accelerate neural networks training from scratch for 3D Zonal Segmentation of the Prostate MRI images.
Conducted extensive evaluation between the proposed UWG-Net with the baseline with small batch sizes, achieving 2-3\% increase in multi-class segmentation accuracy for medical imaging application.
For more info, you can refer to this page !