Semantic segmentation implementation using pytorch framework.
- Fully Convolutional Networks for Semantic Segmentation[paper]
Jonathan Long, Evan Shelhamer, Trevor Darrell - U-Net: Convolutional Networks for Biomedical Image Segmentation [paper]
Olaf Ronneberger, Philipp Fischer, Thomas Brox - Pyramid Scene Parsing Network [paper]
Hengshuang Zhao, Jianping Shi, Xiaojuan Qi, Xiaogang Wang, Jiaya Jia - Learning a Discriminative Feature Network for Semantic Segmentation [paper]
Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, Nong Sang
- FCN8
- U-Net
- PSPNet
- DFNet (CVPR 2018, Implemented part of network)
- I changed some parts of network to lighten network.
- I used augmented dataset. (Currently, I only applied flip operation for augmentation.)
GT images-GT masks-Model predictions
Pascal VOC 2012 / U-Net
Pascal VOC 2012 / FCN8
To train models
python main.py --mode train --model unet --dataset voc \
-- n_iters 10000 --train_batch_size 16 val_batch_size 16 \
--h_image_size 256 --w_image_size 256 \
--model_save_path './models' --sample_save_path './samples'
- Add Cityscape dataset
- Add Semantic boundary dataset
- Add poly learning rate policy
- Change cuda dependency from nsml to local gpu
-
Data argumentation - DeepLab v3 / DeepLab v3+
- Remaining part of DFNet
- Training result
If you have any questions about codes OR find something wrong in my codes, please let me know.