PhysVLM: Vision-Language Model for Generalizable Multitask Remote Physiological Measurement
IEEE Transactions on Instrumentation and Measurement · 2026 · 75 · 1-15
DOI: 10.1109/tim.2026.3671908 ↗At a glance
Journal article in IEEE Transactions on Instrumentation and Measurement, 2026. Written by Jie Gao and 6 co-authors. Cited 1 times so far (Crossref); 47 references. Funded by: Science and Technology Development Fund of Macao, Macao Polytechnic University.
Abstract (original language)
Remote photoplethysmography (rPPG) offers a noninvasive solution for estimating critical vital signs, such as heart rate (HR) and respiratory rate (RR). To improve algorithmic generalizability, domain generalization (DG) has garnered increasing attention for rPPG. However, existing approaches often overlook the frequency periodicity driven by blood-volume pulsations (BVP) and the dependencies among tasks, which limit generalization. To address this, PhysVLM is proposed as a cross-domain, multitask rPPG framework. First, a pretrained vision language model provides frequency-related priors. These priors are injected into a visual-only student network via knowledge distillation. This encourages the student to focus on physiologically driven periodic features. Next, the wavelet temporal block (WTB) is introduced for multiscale time–frequency decomposition and adaptive filtering to suppress task-irrelevant noise. In addition, frequency-aligned spatial attention (FASA) is proposed to perform frequency-aligned bandpass selection within rhythmically prominent regions. Cross-task consistency constraints are incorporated to reduce spectral leakage. Extensive experiments on five public datasets, together with clinical validation on 25 preterm infants in the neonatal intensive care unit (NICU), show that PhysVLM achieves stable improvements in cross-domain evaluations. On the challenging NBHR dataset, the proposed method reduces HR mean absolute error (MAE) from 5.85 to 2.85 bpm. This framework provides a robust foundation for reliable rPPG applications in clinical settings.
Funding
- Science and Technology Development Fund of Macao
- Macao Polytechnic University
Cite
APA
Gao, J., San, I. W., Luo, X., Chen, Z., Yu, Z., Tan, T., & Sun, Y. (2026). PhysVLM: Vision-Language Model for Generalizable Multitask Remote Physiological Measurement. IEEE Transactions on Instrumentation and Measurement, 75, 1-15. https://doi.org/10.1109/tim.2026.3671908
Vancouver
Gao J, San IW, Luo X, Chen Z, Yu Z, Tan T, et al. PhysVLM: Vision-Language Model for Generalizable Multitask Remote Physiological Measurement. IEEE Transactions on Instrumentation and Measurement. 2026;75:1-15. doi:10.1109/tim.2026.3671908
BibTeX
@article{Gao2026,
author = {Gao, Jie and San, Ieong Weng and Luo, Xiangmin and Chen, Zhengxuan and Yu, Zitong and Tan, Tao and Sun, Yue},
title = {PhysVLM: Vision-Language Model for Generalizable Multitask Remote Physiological Measurement},
journal = {IEEE Transactions on Instrumentation and Measurement},
year = {2026},
volume = {75},
pages = {1--15},
doi = {10.1109/tim.2026.3671908}
}
EUVIMED