Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/hasibzunair/masksup-segmentation
[BMVC'2022 Oral] Masked Supervised Learning for Semantic Segmentation.
https://github.com/hasibzunair/masksup-segmentation
Last synced: 4 days ago
JSON representation
[BMVC'2022 Oral] Masked Supervised Learning for Semantic Segmentation.
- Host: GitHub
- URL: https://github.com/hasibzunair/masksup-segmentation
- Owner: hasibzunair
- Created: 2022-02-06T06:12:15.000Z (almost 3 years ago)
- Default Branch: master
- Last Pushed: 2023-02-08T21:31:28.000Z (almost 2 years ago)
- Last Synced: 2023-03-06T17:48:13.517Z (over 1 year ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 12.3 MB
- Stars: 9
- Watchers: 4
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# MaskSup
[](https://huggingface.co/spaces/hasibzunair/masksup-segmentation-demo)
This is official code for our **BMVC 2022 Oral paper**:
[Masked Supervised Learning for Semantic Segmentation](https://arxiv.org/abs/2210.00923)
## 1. Specification of dependencies
This code requires Python 3.8.12. Run the following to install the required packages.
```
conda update conda
conda env create -f environment.yml
conda activate msl
```
## 2a. Get datasets
First, open a folder named
`datasets` in the root folder (`mkdir datasets`). Then, download GLaS, Kvasir & CVC-ClinicDB and NYUDv2 datasets as well as the sribbles from [GitHub Releases](https://github.com/hasibzunair/masksup-segmentation/releases/tag/v1.0). Finally, unzip and move the four folder to `datasets`.
## 2b. Train & Evaluation code
To train and evaluate MaskSup on GLaS or Kvasir & CVC-ClinicDB datasets, you need to change the `EXPERIMENT_NAME` in `trainval_glas_polyp.py` to a name that has glas or polyp. For example to train on GLaS, set `EXPERIMENT_NAME = "glas_masksup"`. Then run:
```
python trainval_glas_polyp.py
```
To train and evaluate MaskSup on NYUDv2 dataset, run:
```
python trainval_nyudv2.py
```
All experiments are conducted on a single NVIDIA 3080Ti GPU. For additional implementation details and results, please refer to the supplementary material [here](https://github.com/hasibzunair/masksup-segmentation/blob/master/media/supplementary_materials.pdf).
## 3. Pre-trained models
We provide pretrained models on [GitHub Releases](https://github.com/hasibzunair/masksup-segmentation/releases/tag/v0.1) for reproducibility.
|Dataset | Backbone | mIoU(%) | Download |
| ---------- | ------- | ------ | -------- |
| GLaS |LeViT-UNet 384 | 76.06 | [download](https://github.com/hasibzunair/masksup-segmentation/releases/download/v0.1/masksupglas76.06iou.pth) |
| Kvasir & CVC-ClinicDB |LeViT-UNet 384 | 84.02 | [download](https://github.com/hasibzunair/masksup-segmentation/releases/download/v0.1/masksuppolyp84.02iou.pth) |
| NYUDv2 |U-Net++ | 39.31 | [download](https://github.com/hasibzunair/masksup-segmentation/releases/download/v0.1/masksupnyu39.31iou.pth) |
## 4. Demo
A HuggingFace Spaces demo of the model trained with MaskSup on NYUDv2 is available at https://huggingface.co/spaces/hasibzunair/masksup-segmentation-demo.
## 5. Citation
```bibtex
@inproceedings{zunair2022masked,
title={Masked Supervised Learning for Semantic Segmentation},
author={Zunair, Hasib and Hamza, A Ben},
booktitle={Proc. British Machine Vision Conference},
year={2022}
}
```
### Acknowledgements
This code base is built on top of the following repositories:
* https://github.com/apple1986/LeViT-UNet
* https://github.com/France1/unet-multiclass-pytorch
* https://github.com/milesial/Pytorch-UNet