https://github.com/prbonn/contmav
[CVPR2024] Open-world Semantic Segmentation Including Class Similarity
https://github.com/prbonn/contmav
Last synced: 9 months ago
JSON representation
[CVPR2024] Open-world Semantic Segmentation Including Class Similarity
- Host: GitHub
- URL: https://github.com/prbonn/contmav
- Owner: PRBonn
- Created: 2024-03-11T17:11:56.000Z (almost 2 years ago)
- Default Branch: master
- Last Pushed: 2025-03-27T09:16:00.000Z (10 months ago)
- Last Synced: 2025-03-27T10:25:50.559Z (10 months ago)
- Language: Python
- Size: 53.2 MB
- Stars: 73
- Watchers: 5
- Forks: 10
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Open-World Semantic Segmentation Including Class Similarity
This is the code repository of the paper Open-World Semantic Segmentation Including Class Similarity, accepted to the IEEE / CVF Computer Vision and Pattern Recognition Conference (CVPR) 2024.
You can find the paper [here](https://www.ipb.uni-bonn.de/wp-content/papercite-data/pdf/sodano2024cvpr.pdf).
## Installation
Install the libraries of the `requirements.yml`, or create a conda environment by `conda env create -f requirements.yml` and then `conda activate openworld`.
The weights of ResNet34 with NonBottleneck 1D block pretrained on ImageNet are available [here](https://drive.google.com/drive/folders/1goULJjHp5-M7nUGlC52uvWaQxn2j3Za1?usp=sharing).
## Training
You can choose your favourite hyperparameters configuration in `args.py`. For training, run
`python train.py --id --dataset_dir --num_classes --batch_size 8`.
The expected data structure is taken from Cityscapes. BDDAnomaly has been converted to Cityscapes format.
## Cite
Please cite us at
```bibtex
@inproceedings{sodano2024cvpr,
author = {Matteo Sodano and Federico Magistri and Lucas Nunes and Jens Behley and Cyrill Stachniss},
title = {{Open-World Semantic Segmentation Including Class Similarity}},
booktitle = {{Proc. of the IEEE/CVF Conf. on Computer Vision and Pattern Recognition (CVPR)}},
year = {2024}
}