NOTE: A fast and flexible implementation based python: https://github.com/lartpang/PySODMetrics
More details can be found at the origin repository: https://github.com/ArcherFMY/sal_eval_toolbox
This forked repository mainly adds the following new features:
- More metrics: E-measure and weighted F-measure.
- Support for drawing according to existing mat files.
- Support custom threshold to binarize masks. Because the masks of some datasets are not exactly a binary map.
- 2020/4/16:
tools/ README.md
- Fixed the description: Plot these value: You can run draw_mat.m draw_twice.m in matlab -> Plot these value: You can run
draw_curves.m
in matlab - Supplement the description of NOTE, about the exported figure.
- Add a piece of content to HOW TO USE.
- Fixed the description: Plot these value: You can run draw_mat.m draw_twice.m in matlab -> Plot these value: You can run
- Delete some useless files.
If you find this code useful in your research, please consider citing:
@article{sal_eval_toolbox,
Author = {Mengyang Feng},
Title = {Evaluation Toolbox for Salient Object Detection.},
Journal = {https://github.com/ArcherFMY/sal_eval_toolbox},
Year = {2018}
}