Research map / Video / Video segmentation and tracking
DeepPhys: Video-Based Physiological Measurement Using Convolutional Attention Networks
Weixuan Chen, Daniel McDuff
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 appearance-guided attention masks to the motion CNN (ablation) | Motion-only CNN · RGB Video I heart rate (participant-dependent, avg of 6 tasks) | MAE (BPM): 1.73 -> 1.50 | Improved |
| add appearance-guided attention masks to the motion CNN (ablation) | Motion-only CNN · RGB Video I heart rate (participant-independent, avg of 6 tasks) | MAE (BPM): 2.63 -> 2.26 | Improved |
| add appearance-guided attention masks (ablation) | Stacked CNN · RGB Video I heart rate (participant-independent, avg of 6 tasks) | MAE (BPM): 2.38 -> 2.26 | Improved |
| add attention masks (ablation) | Motion-only CNN · RGB Video I heart rate task 6 (participant-dependent) | MAE (BPM): 2.95 -> 1.78 | Improved |
| add attention masks (ablation) | Stacked CNN · RGB Video I breathing rate (participant-independent, avg of 6 tasks) | MAE (BPM): 4.57 -> 3.26 | Improved |
| add attention masks (ablation) | Motion-only CNN · RGB Video I breathing rate (participant-dependent, avg of 6 tasks) | MAE (BPM): 3.02 -> 2.86 | Improved |
| stack raw frame with the normalized frame difference as input (ablation) | Motion-only CNN · RGB Video I heart rate (participant-dependent, avg of 6 tasks) | MAE (BPM): 1.73 -> 1.61 | Improved |
| stack raw frame with the normalized frame difference as input (ablation) | Motion-only CNN · RGB Video I breathing rate (participant-independent, avg of 6 tasks) | MAE (BPM): 3.60 -> 4.57 | Hurt |
| add appearance-guided attention masks (ablation) | Motion-only CNN · Infrared Video heart rate (participant-independent) | MAE (BPM): 1.44 -> 0.55 | Improved |
| add appearance-guided attention masks (ablation) | Motion-only CNN · Infrared Video heart rate (participant-independent) | SNR (dB): 9.55 -> 13.2 | Improved |
| train CAN on task 6 (largest head motion) instead of task 1 (ablation) | CAN, participant-independent HR · RGB Video I heart rate (task-independent, avg MAE over 6 test tasks) | MAE (BPM): 5.70 -> 2.15 | Improved |
| train CAN on all tasks instead of only task 6 (ablation) | CAN, participant-independent HR · RGB Video I heart rate (task-independent, avg MAE) | MAE (BPM): 2.15 -> 2.45 | Hurt |
Closest papers by meaning
- The Benefit of Distraction: Denoising Camera-Based Physiological Measurements Using Inverse Attention (ICCV 2021 · similarity 0.86)
- Video-based Remote Physiological Measurement via Cross-verified Feature Disentangling (ECCV 2020 · similarity 0.85)
- PhysFormer++: Facial Video-Based Physiological Measurement with SlowFast Temporal Difference Transformer (IJCV 2023 · similarity 0.85)
- FacePhys: State of the Heart Learning (NeurIPS 2026 · similarity 0.84)
- RhythmMamba: Fast, Lightweight, and Accurate Remote Physiological Measurement (AAAI 2025 · similarity 0.83)
- FactorizePhys: Matrix Factorization for Multidimensional Attention in Remote Physiological Sensing (NeurIPS 2024 · 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.
