Simple k-max deeplab and remax-deeplab implementation using MMDection-3.1

This is our re-implementation of recent panoptic segmentation methods from Google Research: k-max deeplab and remax-deeplab from PKU and S-Lab@NTU.

kMaX-DeepLab: k-means Mask Transformer, ECCV-2022 arxiv

ReMaX: Relaxing for Better Training on Efficient Panoptic Segmentation, NeurIPS-2023 arxiv

Please take a look at the original paper for the details.

Citations

Please consider their works when using this code.

@inproceedings{kmax_deeplab_2022,
  author={Qihang Yu and Huiyu Wang and Siyuan Qiao and Maxwell Collins and Yukun Zhu and Hartwig Adam and Alan Yuille and Liang-Chieh Chen},
  title={{k-means Mask Transformer}},
  booktitle={ECCV},
  year={2022}
}

@article{sun2023remax,
  title={ReMaX: Relaxing for Better Training on Efficient Panoptic Segmentation},
  author={Sun, Shuyang and Wang, Weijun and Yu, Qihang and Howard, Andrew and Torr, Philip and Chen, Liang-Chieh},
  journal={NeurIPS},
  year={2023}
}

License

MIT