Research map / Video / Video segmentation and tracking
Facial Video-based Remote Physiological Measurement via Self-supervised Learning
Zijie Yue, Miaojing Shi, Shuai Ding
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 |
|---|---|---|---|
| use learnable frequency augmentation instead of video shrinking (ablation) | negative sample generation · UBFC-rPPG | MAE: 1.27 -> 0.58 | Improved |
| use learnable frequency augmentation instead of video stretching (interpolation) (ablation) | negative sample generation · UBFC-rPPG | MAE: 0.91 -> 0.58 | Improved |
| use nonlinear FMB (1D RB + BiLSTM) instead of three 1D convs (ablation) | LFA module · UBFC-rPPG | MAE: 1.05 -> 0.58 | Improved |
| use BiLSTM instead of LSTM, GRU or BiGRU in FMB (ablation) | LFA module · UBFC-rPPG | MAE: 0.66, 0.68, 0.60 -> 0.58 | No clear effect |
| add video reconstruction loss (ablation) | LFA module · UBFC-rPPG | MAE: 0.88 -> 0.58 | Improved |
| use three scales instead of one in LFA (ablation) | LFA module · UBFC-rPPG | MAE: 0.83 -> 0.58 | Improved |
| increase number of negative samples k from 1 to 4 (ablation) | LFA module · UBFC-rPPG | MAE: 1.10 -> 0.58 | Improved |
| use local rPPG expert aggregation instead of global estimation (ablation) | rPPG estimator · UBFC-rPPG | MAE: 1.16 -> 0.58 | Improved |
| use region attention instead of 3D ResNet block (ablation) | REA module · UBFC-rPPG | MAE: 1.02 -> 0.58 | Improved |
| use spatio-temporal gating net instead of averaging experts (ablation) | REA module · UBFC-rPPG | MAE: 1.12 -> 0.58 | Improved |
| increase number of experts L from 2 to 9 (ablation) | REA module · UBFC-rPPG | MAE: 0.75 -> 0.58 | Improved |
| add frequency contrastive loss (ablation) | full method · MMVS | MAE: 5.00 -> 2.93 | Improved |
Closest papers by meaning
- PhysFormer: Facial Video-Based Physiological Measurement With Temporal Difference Transformer (CVPR 2022 · similarity 0.94)
- Learning Motion-Robust Remote Photoplethysmography through Arbitrary Resolution Videos (AAAI 2023 · similarity 0.93)
- RhythmMamba: Fast, Lightweight, and Accurate Remote Physiological Measurement (AAAI 2025 · similarity 0.92)
- Remote Photoplethysmography in Real-World and Extreme Lighting Scenarios (CVPR 2025 · similarity 0.91)
- FacePhys: State of the Heart Learning (NeurIPS 2026 · similarity 0.90)
- PhysDiff: Physiology-based Dynamicity Disentangled Diffusion Model for Remote Physiological Measurement (AAAI 2025 · similarity 0.89)
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.
