Research map / Video / Video segmentation and tracking
The Benefit of Distraction: Denoising Camera-Based Physiological Measurements Using Inverse Attention
Ewa M. Nowara, Daniel McDuff, Ashok Veeraraghavan
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 background corruption estimates as LSTM input (ablation) | CAN + BiLSTM denoiser · MMSE-HR | HR MAE (BPM): 2.80 -> 2.27 | Improved |
| add background corruption estimates as LSTM input (ablation) | CAN + BiLSTM denoiser · MR-NIRP (NIR) | HR MAE (BPM): 2.56 -> 2.34 | Improved |
| add background corruption estimates as LSTM input (ablation) | CAN + BiLSTM denoiser · AFRL | HR MAE (BPM): 2.12 -> 2.25 | No clear effect |
| add background corruption estimates as LSTM input (ablation) | CAN + BiLSTM denoiser · AFRL | BR MAE (BPM): 2.49 -> 2.44 | No clear effect |
| add background corruption estimates as LSTM input (ablation) | CAN + BiLSTM denoiser · AFRL Task 1 train, Task 6 test | HR MAE (BPM): 5.61 -> 5.29 | Improved |
| add background corruption estimates as LSTM input (ablation) | CAN + BiLSTM denoiser · AFRL Task 1 train, Task 6 test | BR MAE (BPM): 4.38 -> 4.28 | Improved |
| use learned LSTM denoising instead of frequency-domain subtraction (ablation) | CAN + BiLSTM denoiser · AFRL | HR MAE (BPM): 2.92 -> 2.25 | Improved |
| use learned LSTM denoising instead of time-domain waveform subtraction (ablation) | CAN + BiLSTM denoiser · MMSE-HR | HR MAE (BPM): 6.09 -> 2.27 | Improved |
| use learned LSTM denoising instead of time-domain waveform subtraction (ablation) | CAN + BiLSTM denoiser · MR-NIRP (NIR) | HR MAE (BPM): 8.83 -> 2.34 | Improved |
| use distraction regions near the face instead of image edges (ablation) | CAN + BiLSTM denoiser · AFRL large head motion | HR MAE (BPM): 8.74 -> 6.53 | Improved |
| use distraction regions near the face instead of image edges (ablation) | CAN + BiLSTM denoiser · AFRL small motion | HR MAE (BPM): 1.07 -> 1.08 | No clear effect |
| continuous vs binary inverse mask, all vs single RGB channels (ablation) | CAN + BiLSTM denoiser · AFRL | HR MAE (BPM) | No clear effect |
Closest papers by meaning
- Contrast-Phys+: Unsupervised and Weakly-Supervised Video-Based Remote Physiological Measurement via Spatiotemporal Contrast (TPAMI 2024 · similarity 0.86)
- DeepPhys: Video-Based Physiological Measurement Using Convolutional Attention Networks (ECCV 2018 · similarity 0.86)
- FactorizePhys: Matrix Factorization for Multidimensional Attention in Remote Physiological Sensing (NeurIPS 2024 · similarity 0.85)
- Faster Video Diffusion with Trainable Sparse Attention (NeurIPS 2025 · similarity 0.83)
- Mind the Gap: a Spectral Analysis of Rank Collapse and Signal Propagation in Attention Layers (ICML 2025 · similarity 0.83)
- 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.
