Our motivation is to optimize current models on accuracy and robustness regarding lumbar spine images.
This project contains Pytorch implementation of various basic models.
- UNet proposed in the paper U-Net: Convolutional Networks for Biomedical Image Segmentation
- Python 3.7
- Pytorch 1.2
- torchsummary
- tb-nightly
$ python train.py
-
Generated samples with combined losses(100th epochs)
-
The loss of training/validing dataset