rPPG-VQA filters in-the-wild videos using signal-level SNR consensus and scene-level MLLM interference detection, then applies two-stage adaptive sampling to produce unsupervised rPPG models with substantially higher benchmark accuracy.
Remote heart rate measurement from highly compressed facial videos: an end-to-end deep learning so- lution with video enhancement
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
rPPG-VQA: A Video Quality Assessment Framework for Unsupervised rPPG Training
rPPG-VQA filters in-the-wild videos using signal-level SNR consensus and scene-level MLLM interference detection, then applies two-stage adaptive sampling to produce unsupervised rPPG models with substantially higher benchmark accuracy.