Explanations for Occluded Images
Abstract
Existing algorithms for explaining the output of image classifiers perform poorly on inputs where the object of interest is partially occluded. We present a novel, black-box algorithm for computing explanations that uses a principled approach based on causal theory. We have implemented the method in the DeepCover tool. We obtain explanations that are much more accurate than those generated by the existing explanation tools on images with occlusions and observe a level of performance comparable to the state of the art when explaining images without occlusions.
Cite
Text
Chockler et al. "Explanations for Occluded Images." International Conference on Computer Vision, 2021. doi:10.1109/ICCV48922.2021.00127Markdown
[Chockler et al. "Explanations for Occluded Images." International Conference on Computer Vision, 2021.](https://mlanthology.org/iccv/2021/chockler2021iccv-explanations/) doi:10.1109/ICCV48922.2021.00127BibTeX
@inproceedings{chockler2021iccv-explanations,
title = {{Explanations for Occluded Images}},
author = {Chockler, Hana and Kroening, Daniel and Sun, Youcheng},
booktitle = {International Conference on Computer Vision},
year = {2021},
pages = {1234-1243},
doi = {10.1109/ICCV48922.2021.00127},
url = {https://mlanthology.org/iccv/2021/chockler2021iccv-explanations/}
}