Research map / Adversarial robustness and security / Black-box and transfer attacks
Cross-Modal Transferable Adversarial Attacks From Images to Videos
Zhipeng Wei, Jingjing Chen, Zuxuan Wu, Yu-Gang Jiang
Check 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 |
|---|---|---|---|
| Minimize cosine similarity between benign and adversarial intermediate image features (I2V) rather than use DR | Cross-modal image-to-video attack · UCF-101 | AASR: +63.33% | Improved |
| Minimize cosine similarity between benign and adversarial intermediate image features (I2V) rather than use DR | Cross-modal image-to-video attack · Kinetics-400 | AASR: +42.51% | Improved |
| Use an ensemble of ImageNet-pretrained image models instead of a single image model | I2V · UCF-101 | AASR: 65.68% | Improved |
| Use an ensemble of ImageNet-pretrained image models instead of a single image model | I2V · Kinetics-400 | AASR: 77.88% | Improved |
| Use Resnet-101 as the image model rather than other tested single image models | I2V · UCF-101 and Kinetics-400 | AASR | Improved |
| Attack middle image-model layers rather than bottom or top layers (ablation) | I2V · Kinetics-400 | AASR | Improved |
| Use smaller step size and fewer optimization iterations (ablation) | I2V · Kinetics-400 | AASR | Hurt |
| Use larger step size (ablation) | I2V · Kinetics-400 | AASR | Improved |
| Use I2V instead of ENS-I2V | Cross-modal attack on video models · Kinetics-400, TPN-101 and TPN-50 | ASR | Mixed |
| Use image-model-based I2V instead of video-model-based ILAF | Transfer-based attack · Kinetics-400 with TPN-101 as white-box model | AASR | Hurt |
Closest papers by meaning
- Adaptive Cross-Modal Transferable Adversarial Attacks From Images to Videos (TPAMI 2023 · similarity 0.99)
- Global-Local Characteristic Excited Cross-Modal Attacks from Images to Videos (AAAI 2023 · similarity 0.94)
- Transferability of Adversarial Attacks in Video-based MLLMs: A Cross-modal Image-to-Video Approach (AAAI 2026 · similarity 0.91)
- Boosting the Transferability of Video Adversarial Examples via Temporal Translation (AAAI 2022 · similarity 0.90)
- Cross-Domain Transferability of Adversarial Perturbations (NeurIPS 2019 · similarity 0.89)
- Enhancing Adversarial Transferability with Adversarial Weight Tuning (AAAI 2025 · similarity 0.87)
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.
