Research map / 3D vision / Depth and stereo
CliffPhys: Camera-based Respiratory Measurement using Clifford Neural Networks
Omar Ghezzi, Giuseppe Boccignone, Giuliano Grossi, Raffaella Lanzarotti, Alessandro D'Amelio
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 monocular depth to the optical flow input (ablation) | CliffPhys Cl0,2 · COHFACE test | MAE (breaths/min): 0.90 -> 0.83 | Improved |
| add monocular depth to the optical flow input (ablation) | CliffPhys Cl3,0 · BP4D+ cross-dataset | MAE (breaths/min): 4.23 -> 3.50 | Improved |
| add depth to input for the simpler Cl2,0 algebra (ablation) | CliffPhys Cl2,0 · COHFACE test | MAE (breaths/min): 0.95 -> 1.06 | Hurt |
| use optical flow and depth input instead of raw RGB frames (ablation) | CliffPhys Cl3,0 · COHFACE test | MAE (breaths/min): 4.15 -> 0.99 | Improved |
| use Clifford layers instead of conventional 3D convolutional layers on the same flow and depth input (ablation) | CliffPhys · COHFACE test | MAE (breaths/min): 3.62 -> 0.83 | Improved |
| use 8-dimensional Clifford algebra (Cl3,0, Cl0,3) instead of 4-dimensional on cross-dataset (ablation) | CliffPhys · BP4D+ cross-dataset | MAE (breaths/min): Cl0,2 4.00 and Cl2,0 4.07 -> Cl3,0 3.50 | Improved |
| use CliffPhys instead of RGB-based neural baselines (MTTS-CAN, BigSmall, PhysFormer, ContrastPhys) | CliffPhys · COHFACE test | MAE (breaths/min): 2.83 (MTTS-CAN) -> 0.83 | Improved |
| compare CliffPhys Cl3,0 to ContrastPhys on cross-dataset MAE | CliffPhys · BP4D+ cross-dataset | MAE (breaths/min): 3.82 -> 3.50 | Improved |
Closest papers by meaning
- DeepPhys: Video-Based Physiological Measurement Using Convolutional Attention Networks (ECCV 2018 · similarity 0.82)
- FacePhys: State of the Heart Learning (NeurIPS 2026 · similarity 0.81)
- SCAMPS: Synthetics for Camera Measurement of Physiological Signals (NeurIPS 2022 · similarity 0.80)
- Fusion-Vital: Video-RF Fusion Transformer for Advanced Remote Physiological Measurement (AAAI 2024 · similarity 0.79)
- Towards Open Respiratory Acoustic Foundation Models: Pretraining and Benchmarking (NeurIPS 2024 · similarity 0.79)
- PhysDiff: Physiology-based Dynamicity Disentangled Diffusion Model for Remote Physiological Measurement (AAAI 2025 · similarity 0.78)
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.
