Vision4PPG: Emergent PPG Analysis Capability of Vision Foundation Models for Vital Signs like Blood Pressure

Vision4PPG turns 1D PPG into 2D images (STFT magnitude, STFT phase, and recurrence plots) and fine-tunes public vision foundation models, DINOv3 and SigLIP-2 ViT-B (86M parameters), with LoRA. It is an adaptation of vision models, not a PPG-pretrained foundation model. In the authors' Table 1, those vision models win 9 of 14 diastolic/systolic MAE comparisons against fully tuned MOMENT-385M and PPG-GPT-345M; SigLIP-2's Aurora-Oscillometric DBP/SBP MAE is 5.32/18.57 mmHg. The bed BCG set is an exception, where MOMENT and PPG-GPT report lower error than either vision model.

• Authors: Kataria, Ermis, Panchumarthi, Wang, Hu (Emory Nursing)
• arXiv:2510.10366v1 (11 Oct 2025); DOI 10.48550/arXiv.2510.10366
• DINOv3 and SigLIP-2 ViT-B, 86M; LoRA on Q, K, V (r=8, alpha=16, dropout 0.05)
• Table 1: SigLIP-2 Aurora-Oscillometric DBP/SBP MAE 5.32/18.57 mmHg
• Authors: vision FMs win 9/14 DBP/SBP comparisons vs MOMENT-385M and PPG-GPT-345M
• Table 2: PPG-GPT HR MAE 7.83 bpm (DaLiA) beats both vision models; DINOv3 respiration MAE 4.41 breaths/min (BIDMC) and SpO2 MAE 2.19% (MIMIC-III)

Photoplethysmography (PPG) sensor in wearable and clinical devices provides valuable physiological insights in a non-invasive and real-time fashion. Specialized Foundation Models (FM) or repurposed time-series FMs are used to benchmark physiological tasks. Our experiments with fine-tuning FMs reveal that Vision FM (VFM) can also be utilized for this purpose and, in fact, surprisingly leads to state-of-the-art (SOTA) performance on many tasks, notably blood pressure estimation. We leverage VFMs by simply transforming one-dimensional PPG signals into image-like two-dimensional representations, such as the Short-Time Fourier transform (STFT). Using the latest VFMs, such as DINOv3 and SIGLIP-2, we achieve promising performance on other vital signs and blood lab measurement tasks as well. Our proposal, Vision4PPG, unlocks a new class of FMs to achieve SOTA performance with notable generalization to other 2D input representations, including STFT phase and recurrence plots. Our work improves upon prior investigations of vision models for PPG by conducting a comprehensive study, comparing them to state-of-the-art time-series FMs, and demonstrating the general PPG processing ability by reporting results on six additional tasks. Thus, we provide clinician-scientists with a new set of powerful tools that is also computationally efficient, thanks to Parameter-Efficient Fine-Tuning (PEFT) techniques.

Verbatim from arXiv (opens in a new tab) · arXiv metadata (abstracts are released under CC0)

No results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.

• PPG-BP Database (Guilin / Liang 2018)219 subjects · PPG
• Brno University of Technology Smartphone PPG Database (BUT PPG)50 subjects · PPG · ECG
• VitalVideosface video + contact PPG + BP/SpO2 (Europe-1, Africa-1, Worldwide)850 subjects · PPG
• PPG-DaLiAPPG dataset for motion compensation / HR15 subjects · PPG · ECG · IMU
• BIDMC PPG and Respiration (Pimentel 2016 / PhysioNet)53 subjects · PPG · ECG · SpO₂
• Bed-Based BCG (Carlson, K-State)40 adults, bed BCG + ECG + PPG + Finometer PRO continuous BP (IEEE DataPort OA)40 subjects · PPG · ECG · BP

• Vision4PPGtransformerPPG2025