# Why Do VLMs Fail at Remote Physiological Measurement? Rethinking with Physiological Grounding and Numerical Alignment

arXiv: https://arxiv.org/abs/2610.19052 · HTML: https://mlsanity.com/paper/2610.19052/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 19 of its 30 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Placed in

- LLM reasoning
- Multimodal LLMs / vision-language models

## Builds on (most cited first, 19 of 19)

- [Video-LLaVA: Learning United Visual Representation by Alignment Before Projection](https://mlsanity.com/paper/2311.10122/) (2023)
- [BLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and Dataset](https://mlsanity.com/paper/2505.09568/) (2025)
- [Remote Photoplethysmograph Signal Measurement from Facial Videos Using Spatio-Temporal Networks](https://mlsanity.com/paper/1905.02419/) (2019)
- [PhysFormer: Facial Video-based Physiological Measurement with Temporal Difference Transformer](https://mlsanity.com/paper/2111.12082/) (2021)
- [RhythmNet: End-to-end Heart Rate Estimation from Face via Spatial-temporal Representation](https://mlsanity.com/paper/1910.11515/) (2019)
- [VIPL-HR: A Multi-modal Database for Pulse Estimation from Less-constrained Face Video](https://mlsanity.com/paper/1810.04927/) (2018)
- [MMPD: Multi-Domain Mobile Video Physiology Dataset](https://mlsanity.com/paper/2302.03840/) (2023)
- [Neuron Structure Modeling for Generalizable Remote Physiological Measurement](https://mlsanity.com/paper/2303.05955/) (2023)
- [RhythmFormer: Extracting Patterned rPPG Signals based on Periodic Sparse Attention](https://mlsanity.com/paper/2402.12788/) (2024)
- [Image Enhancement for Remote Photoplethysmography in a Low-Light Environment](https://mlsanity.com/paper/2303.09336/) (2023)
- [FactorizePhys: Matrix Factorization for Multidimensional Attention in Remote Physiological Sensing](https://mlsanity.com/paper/2411.01542/) (2024)
- [PhysLLM: Harnessing Large Language Models for Cross-Modal Remote Physiological Sensing](https://mlsanity.com/paper/2505.03621/) (2025)
- [Regress, Don't Guess -- A Regression-like Loss on Number Tokens for Language Models](https://mlsanity.com/paper/2411.02083/) (2024)
- [Efficient Mixture-of-Expert for Video-based Driver State and Physiological Multi-task Estimation in Conditional Autonomous Driving](https://mlsanity.com/paper/2410.21086/) (2024)
- [PhysDrive: A Multimodal Remote Physiological Measurement Dataset for In-vehicle Driver Monitoring](https://mlsanity.com/paper/2507.19172/) (2025)
- [Not Only Consistency: Enhance Test-Time Adaptation with Spatio-temporal Inconsistency for Remote Physiological Measurement](https://mlsanity.com/paper/2507.07908/) (2025)
- [Passive Heart Rate Monitoring During Smartphone Use in Everyday Life](https://mlsanity.com/paper/2503.03783/) (2025)
- [MILD: Mediator Agent System with Bidirectional Perception and Multi-Layered Alignment for Human-Vehicle Collaboration](https://mlsanity.com/paper/2605.01507/) (2026)
- [RPM-Distill: Physiology-guided Adaptive Cross-modal Distillation for Robust Remote Physiological Measurement](https://mlsanity.com/paper/2606.28089/) (2026)

## Abstract (arXiv)

Video-based remote physiological measurement (RPM) has broad applications, while Vision-Language Models (VLMs) provide a promising opportunity to enable end-to-end physiological monitoring for future human-aware intelligent systems. However, existing RPM methods rely on task-specific models and cannot be seamlessly integrated with the multimodal understanding and interaction capabilities of VLMs. We revisit RPM with VLMs and provide empirical evidence for two major bottlenecks: generic visual semantic representations fail to preserve fine-grained physiological cues; while discrete token prediction in autoregressive generation is inherently misaligned with continuous numerical measurement. To address these challenges, we propose VITAL (Visual-physiological Token Alignment for Language Models), a visual-physiological token alignment framework for RPM and preserving general vision tasks. VITAL retains the original visual pathway for semantic understanding and introduces an independent physiological pathway to jointly model semantic and physiological information. Moreover, we propose a structured numerical alignment approach to improve the reliability of generative heart rate estimation by enhancing numerical relationship modeling. Extensive experiments show that VITAL improves cross-domain estimation performance on multiple rPPG datasets while preserving the general visual understanding capability of VLMs, demonstrating its effectiveness as a unified visual-physiological model. Code will be released.