Research map / Adversarial robustness and security / Attacks on text and graphs
Oulu Remote-photoplethysmography Physical Domain Attacks Database (ORPDAD)
Marko Savic, Guoying Zhao
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 physical attacks to the training data (attack aware training) (ablation) | supervised rPPG methods (DeepPhys, TS-CAN, PhysNet) · ORPDAD illumination I3+I6 | RMSE: PhysNet 16.1 -> 1.61; DeepPhys 26.8 -> 13.9 | Improved |
| attack aware training for self-supervised Contrast-Phys (ablation) | Contrast-Phys · ORPDAD all categories | RMSE: still 0.95 -> 20.1; I3+I6 17.2 -> 21.9 | Hurt |
| use non-end-to-end spatial-temporal map input (CVD, PhySU-Net) instead of end-to-end input | supervised rPPG methods · ORPDAD illumination attacks, attack unaware | RMSE: I3+I6: PhySU-Net 6.24, CVD 7.17 vs TS-CAN 27.4 | Improved |
| non-end-to-end methods under movement attacks versus illumination | CVD, PhySU-Net · ORPDAD movement attacks | RMSE | Mixed |
| use physiology-based model methods (CHROM, POS) instead of GREEN, ICA, PCA | hand-crafted rPPG · ORPDAD illumination and movement attacks | RMSE: I3+I6: POS 13.8, CHROM 19.5 vs GREEN 27.7 | Improved |
| increase illumination attack intensity from small to large (ablation) | hand-crafted and supervised rPPG methods · ORPDAD illumination attacks | RMSE: CHROM 1.87 -> 19.5 | Hurt |
| periodic mouth movement attack versus natural talking (ablation) | hand-crafted rPPG · ORPDAD movement | RMSE: CHROM M5+M10 10.6 vs M11 1.35 | Hurt |
| cover face with mask (C2) versus forehead hat (C1) (ablation) | ROI-based methods · ORPDAD concealment | RMSE: CHROM C1 0.72 -> C2 11.1 | Hurt |
| apply foundation versus transparent primer (ablation) | rPPG methods · ORPDAD concealment C4-C6 | SNR: about 1 dB drop | Hurt |
| add setting powder layer over foundation | rPPG methods · ORPDAD concealment C5 vs C6 | SNR | No clear effect |
Closest papers by meaning
- Remote Photoplethysmography Correspondence Feature for 3D Mask Face Presentation Attack Detection (ECCV 2018 · similarity 0.89)
- Dual-Bridging With Adversarial Noise Generation for Domain Adaptive rPPG Estimation (CVPR 2023 · similarity 0.87)
- Remote Photoplethysmography in Real-World and Extreme Lighting Scenarios (CVPR 2025 · similarity 0.87)
- Augmentation of rPPG Benchmark Datasets: Learning to Remove and Embed rPPG Signals via Double Cycle Consistent Learning from Unpaired Facial Videos (ECCV 2022 · similarity 0.87)
- RhythmMamba: Fast, Lightweight, and Accurate Remote Physiological Measurement (AAAI 2025 · similarity 0.87)
- PhysDiff: Physiology-based Dynamicity Disentangled Diffusion Model for Remote Physiological Measurement (AAAI 2025 · similarity 0.83)
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.
