方法对比

	可解释实例分割	可解释视觉 Transformer
领域	深度学习	深度学习
方法族	Machine learning	Machine learning
起源年份	2017–present	2021
提出者	He, K. et al. (Mask R-CNN); XAI extensions by multiple authors	Chefer, H., Gur, S., & Wolf, L. (attribution framework); Dosovitskiy et al. (base ViT)
类型	Explainability-augmented deep learning pipeline	Post-hoc explainability applied to Vision Transformer
开创性文献	Lindner, M., Meng, C., & Bischl, B. (2023). Explaining Instance Segmentation Models via Saliency Maps and Occlusion. IEEE Transactions on Pattern Analysis and Machine Intelligence.	Chefer, H., Gur, S., & Wolf, L. (2021). Transformer interpretability beyond attention visualization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 782–791.
别名	XAI instance segmentation, interpretable instance segmentation, transparent mask prediction, explainable Mask R-CNN	XViT, Interpretable ViT, Explainable ViT, Transparent Vision Transformer
相关	6	5
摘要	Explainable Instance Segmentation combines deep-learning instance segmentation models — which detect and delineate every individual object as a separate pixel mask — with post-hoc or ante-hoc explainability techniques such as GradCAM, SHAP, LIME, or attention visualization, so that each predicted mask is accompanied by evidence showing which image regions drove the model's decision.	Explainable Vision Transformer combines the strong image-recognition performance of Vision Transformers (ViT) with attribution techniques — such as relevance propagation, attention rollout, or gradient-weighted attention — that highlight which image regions drive each prediction. The approach enables researchers and practitioners to audit model decisions and satisfy transparency requirements without sacrificing accuracy.