Full-Self Diagnostics (FSD): Physics-Grounded Visual Biomarker Inference from Smartphone Video via Inverse Problems and Operator Learning

2026-06-11

AI Summary

This work proposes a non-invasive method for estimating clinical physiological biomarkers, such as blood glucose levels, from unconstrained 9-second facial videos captured by standard smartphones. The approach is grounded in a unified mathematical framework that integrates a radiative transfer-based physical forward model with multimodal visual signals—including spectral, pulse, respiratory, micro-expression, and eye-movement cues. By combining Tikhonov-regularized inverse problem solving with operator learning, and leveraging information-theoretic observability theory to demonstrate the informational gain of multimodal signals for physiological state inference, the system achieves robust performance. Validated on 38,812 real-world paired samples, it attains an overall MARD of 29.86%, with 97.57% of predictions falling within Clarke Error Grid zones A+B; notably, MARD improves to 17% in the clinically critical 70–180 mg/dL range. The method exhibits strong generalizability across devices and populations, along with interpretability and data-driven performance enhancement.

Abstract

We present Full-Self Diagnostics (FSD), a unified mathematical framework for recovering latent physiological states from unconstrained 9-second facial videos captured by consumer smartphones. The approach integrates five mutually reinforcing components: (1) a physics-based forward model derived from the radiative transfer equation and chromophore absorption that maps camera observables to biomarker concentrations; (2) an information-theoretic observability theory proving that multi-channel visual signals (spectral, pulse, respiratory, micro-expression, and oculomotor) contain strictly increasing mutual information with physiological state; (3) a stable, Tikhonov-regularized inverse problem with domain-uniform identifiability guarantees; (4) an operator-learning formulation that enables generalization across devices, resolutions, and populations; and (5) a supervised learning procedure, interpretable as stochastic variational inference, that continuously refines the model from paired biosensor ground truth with performance improving proportionally to one over the square root of the number of paired observations. Empirical validation on 38812 real-world paired scans across 59 subjects demonstrates practical performance. Self-collected data from the lead author (glucose range 35-550 mg/dL) yields MARD of 29.86 percent with 97.57 percent of predictions in Clarke Error Grid Zones A+B and only 0.27 percent in the dangerous Zone E. A well-managed diabetic participant achieves MARD of 17 percent in the narrower 70-180 mg/dL band. These results confirm that consumer-grade facial video encodes sufficient structured information for clinically relevant, non-invasive biomarker inference under fully unconstrained conditions, with performance scaling predictably as more paired data becomes available.

Problem

• biomarker inference
• smartphone video
• non-invasive diagnostics
• physiological state estimation
• visual biomarkers

Innovation

• physics-informed inverse problems
• operator learning
• visual biomarker inference
• smartphone-based diagnostics
• information-theoretic observability