Research map / Video / Video segmentation and tracking
Spatial Alignment and Temporal Matching Adapter for Video-Radar Remote Physiological Measurement
Qian Liang, Ruixu Geng, Jinbo Chen, Haoyu Wang, Yan Chen, Yang Hu
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 Spatial Alignment to the vanilla adapter (ablation) | SATM adapter · MMRPM | RMSE (bpm): 5.453 -> 5.304 | Improved |
| add Temporal Matching on top of Spatial Alignment (ablation) | SATM adapter · MMRPM | MAE (bpm): 1.454 -> 1.292 | Improved |
| use cross-attention in time-series space for spatial alignment instead of interpolation (ablation) | Spatial Alignment · MMRPM | MAE (bpm): 1.876 -> 1.454 | Improved |
| align in time-series space instead of flattening channel and time (ablation) | Spatial Alignment · MMRPM | MAE (bpm): 1.693 -> 1.454 | Improved |
| use matched-filter Temporal Matching instead of temporal cross-attention (ablation) | SATM adapter · MMRPM | MAE (bpm): 1.367 -> 1.292 | Improved |
| insert SATM adapters only in the last two stages (SATM-2) instead of one, three, or four (ablation) | SATM adapter · MMRPM | MAE (bpm): 1.614 (SATM-1), 1.343 (SATM-3), 1.386 (SATM-4) vs 1.292 (SATM-2) | Mixed |
| add unimodal pre-training to existing video-radar methods | CardiacMamba · EquiPleth | RMSE (bpm): 2.927 -> 2.771 | Improved |
| add unimodal pre-training to Fusion-Vital | Fusion-Vital · EquiPleth | MAE (bpm): 3.750 -> 3.750 | No clear effect |
| replace full fine-tuning fusion with SATM adapters on frozen pre-trained backbones | EquiPleth* · EquiPleth | RMSE (bpm): 1.436 -> 1.242 | Improved |
| SATM vs LAVISH adapter in cross-dataset transfer | LAVISH · EquiPleth to MMRPM | RMSE (bpm): 8.232 -> 7.656 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Fusion-Vital* — Fusion-Vital: Video-RF Fusion Transformer for Advanced Remote Physiological Measurement (AAAI 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- RPM-Distill: Physiology-guided Adaptive Cross-modal Distillation for Robust Remote Physiological Measurement (ECCV 2026 · similarity 0.90)
- FusionPhys: A Flexible Framework for Fusing Complementary Sensing Modalities in Remote Physiological Measurement (ICCV 2025 · similarity 0.87)
- Video-based Remote Physiological Measurement via Cross-verified Feature Disentangling (ECCV 2020 · similarity 0.87)
- RhythmMamba: Fast, Lightweight, and Accurate Remote Physiological Measurement (AAAI 2025 · similarity 0.86)
- Fusion-Vital: Video-RF Fusion Transformer for Advanced Remote Physiological Measurement (AAAI 2024 · similarity 0.85)
- PhysDrive: A Multimodal Remote Physiological Measurement Dataset for In-vehicle Driver Monitoring (NeurIPS 2025 · similarity 0.84)
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.
