Research map / Video / Action recognition
Back to the Feature: Explaining Video Classifiers with Video Counterfactual Explanations
Chao Wang, Chengan Che, Xinyue Chen, Sophia Tsoka, Luis C. Garcia-Peraza-Herrera
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add inversion stage before counterfactual optimization (ablation) | BTTF · NTU RGB+D (A-swinR) | proximity (qualitative) | Improved |
| add video style loss regularization (ablation) | BTTF · NTU RGB+D (A-swinR) | visual realism (qualitative) | Improved |
| replace classifier-guidance video baseline CG-Video with BTTF | BTTF · Shape-Moving (M-swin) | FID: 68.01 -> 27.49 | Improved |
| replace classifier-guidance video baseline CG-Video with BTTF | BTTF · Shape-Moving (M-swin) | FVD: 682.23 -> 250.11 | Improved |
| use video-level guidance with full-noise start instead of half-noisy start | CG-Video vs CG-Video-Mid · Shape-Moving (M-swin) | FVD: 1796.21 -> 682.23 | Improved |
| use video-classifier gradients instead of per-frame guidance | CG-Video-Mid vs CG-Frame · Shape-Moving (M-swin) | flip rate: 0.28 -> 1.00 | Improved |
| use BTTF instead of PGD Attack on non-robust classifier | BTTF · MEAD (E-swin) | FID: 0.68 -> 19.03 | Hurt |
| use BTTF instead of PGD Attack on non-robust classifier | BTTF · MEAD (E-swin) | FVD: 0.53 -> 275.44 | Hurt |
| use BTTF instead of PGD Attack on non-robust classifier | BTTF · MEAD (E-swin) | SSIM: 0.99 -> 0.81 | Hurt |
| use BTTF instead of CG-Video | BTTF · Shape-Moving (M-swin) | flip rate: 1.00 -> 0.97 | No clear effect |
Closest papers by meaning
- Rethinking Visual Counterfactual Explanations Through Region Constraint (ICLR 2025 · similarity 0.87)
- Towards Fine-Grained Interpretability: Counterfactual Explanations for Misclassification with Saliency Partition (CVPR 2025 · similarity 0.86)
- Counterfactual Explanations for Misclassified Images: How Human and Machine Explanations Differ (Abstract Reprint) (AAAI 2024 · similarity 0.86)
- CounterFlowNet: From Minimal Changes to Meaningful Counterfactual Explanations (NeurIPS 2026 · similarity 0.85)
- V-CECE: Visual Counterfactual Explanations via Conceptual Edits (NeurIPS 2025 · similarity 0.85)
- LeapFactual: Reliable Visual Counterfactual Explanation Using Conditional Flow Matching (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
