Modified and Enhanced on excellent work by https://github.com/saic-vul/fbrs_interactive_segmentation/
- Added Selection of different classes
- Save outputs to json file on keystroke ctrl+s
- The output json file can be used directly to train detectron2
Example usage:
python demo.py --gpu 0 --checkpoint hrnet32_ocr128_lvis.pth
@article{fbrs2020,
title={f-BRS: Rethinking Backpropagating Refinement for Interactive Segmentation},
author={Konstantin Sofiiuk, Ilia Petrov, Olga Barinova, Anton Konushin},
journal={arXiv preprint arXiv:2001.10331},
year={2020}
}