AI Paper Atlas

Research map / Video / Video segmentation and tracking

Contrast-Phys+: Unsupervised and Weakly-Supervised Video-Based Remote Physiological Measurement via Spatiotemporal Contrast

Zhaodong Sun, Xiaobai Li

TPAMI 2024 · Video › Video segmentation and tracking · approach: Sign language and physiological video analysis · 81 citations (OpenAlex)

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 changedWhereMeasuredEffect
add HR range constraint (ablation)Contrast-Phys+ (0%) · UBFC-rPPGMAE (bpm): 39.66 -> 0.64Improved
add rPPG cross-video dissimilarity (negative loss) (ablation)Contrast-Phys+ (0%) · UBFC-rPPGMAE (bpm): 22.11 -> 0.64Improved
add rPPG temporal similarity (ablation)Contrast-Phys+ (0%) · UBFC-rPPGMAE (bpm): 1.26 -> 0.64Improved
add rPPG spatial similarity (2x2 instead of 1x1 resolution) (ablation)Contrast-Phys+ (0%) · UBFC-rPPGMAE (bpm): 3.14 -> 0.64Improved
use 10s ST-rPPG time length instead of 30s (ablation)Contrast-Phys+ (0%) · UBFC-rPPGMAE (bpm): 1.97 -> 0.64Improved
use rPPG sample length T/2 instead of T/4 (ablation)Contrast-Phys+ (0%) · UBFC-rPPGMAE (bpm): 0.98 -> 0.64Improved
add GT-rPPG positive loss (ablation)Contrast-Phys+ (100%) · MMSE-HRMAE (bpm): 1.82 -> 1.39Improved
add GT-rPPG negative loss (ablation)Contrast-Phys+ (100%) · MMSE-HRMAE (bpm): 1.82 -> 1.77Improved
add both GT-related positive and negative losses (ablation)Contrast-Phys+ (100%) · MMSE-HRMAE (bpm): 1.82 -> 1.11Improved
add unlabeled PURE videos to labeled UBFC trainingContrast-Phys+ (100%) · MMSE-HR (cross-dataset)MAE (bpm): 1.76 -> 1.13Improved
add unlabeled OBF videos to labeled UBFC trainingContrast-Phys+ (100%) · MMSE-HR (cross-dataset)MAE (bpm): 1.76 -> 1.47Improved
add unlabeled MMSE-HR to labeled SCAMPS trainingContrast-Phys+ · UBFC-rPPGMAE (bpm): 0.89 -> 0.51Improved

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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy