https://github.com/Jerry00917/samshap
Code for the paper "Explain Any Concept: Segment Anything Meets Concept-Based Explanation". Poster @ NeurIPS 2023
https://github.com/Jerry00917/samshap
Last synced: about 1 month ago
JSON representation
Code for the paper "Explain Any Concept: Segment Anything Meets Concept-Based Explanation". Poster @ NeurIPS 2023
- Host: GitHub
- URL: https://github.com/Jerry00917/samshap
- Owner: Jerry00917
- License: other
- Created: 2023-10-14T03:35:41.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2023-12-04T05:39:21.000Z (over 1 year ago)
- Last Synced: 2024-10-27T07:32:22.390Z (6 months ago)
- Language: Python
- Homepage:
- Size: 1.23 MB
- Stars: 40
- Watchers: 2
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Explain Any Concept: Segment Anything Meets Concept-Based Explanation (EAC) Poster @ NeurIPS 2023
Code for the paper "Explain Any Concept: Segment Anything Meets Concept-Based Explanation".
## Citation
Please cite the paper as follows if you use the data or code from Samshap:
```
@inproceedings{
sun2023explain,
title={Explain Any Concept: Segment Anything Meets Concept-Based Explanation},
author={Ao Sun and Pingchuan Ma and Yuanyuan Yuan and Shuai Wang},
booktitle={Thirty-seventh Conference on Neural Information Processing Systems},
year={2023},
url={https://openreview.net/forum?id=X6TBBsz9qi}
}
```
## Contact
Please reach out to us if you have any questions or suggestions. You can send an email to [email protected].
## Overview
Here is an overview of our work, and you can find more in our [Paper](https://openreview.net/forum?id=X6TBBsz9qi).
Our EAC approach generates high accurate and human-understandable post-hoc explanations.
## Downloading the SAM backbone
We use ViT-H as our default SAM model. For downloading the pre-train model and installation dependencies, please refer [SAM repo](https://github.com/facebookresearch/segment-anything#model-checkpoints).
## Explain a hummingbird on your local pre-trained ResNet-50!
Simply run the following command:
```
python demo_samshap.py
```