Representation Is Not Enough: Body-Localized Thermal Evidence for Contactless Stress and Craving Sensing in Opioid Use Disorder

2026-08-17

AI Summary

This study addresses signal localization and equitable deployment challenges in contactless monitoring for opioid use disorder (OUD) by proposing FABLE-Therm. This architecture integrates frozen foundation model encoders while preserving localized thermal evidence to enable weakly supervised stress and craving detection alongside failure attribution. The work demonstrates for the first time that contactless thermal video can recover craving states, revealing that retaining local evidence outperforms feature concatenation and that improving representations alone is insufficient to ensure fairness. Achieving an AUROC of 0.938 on an independent cohort, this research establishes the first contactless thermal imaging benchmark for OUD and introduces a comprehensive framework for analyzing deployment failures, thereby advancing both technical performance and responsible implementation in clinical monitoring.

Abstract

Removing wearables from physiological monitoring also removes their supervision: the signal indicating where and when a stress response occurred. Contactless stress sensing therefore becomes a weakly supervised evidence-localization problem, where a clip-level label must be traced to the body regions and moments that produced it. We address this with FABLE-Therm, a weakly supervised architecture that preserves localized evidence across body regions, time, and encoder-specific representations until the final decision. FABLE-Therm fuses frozen foundation-model encoders at the embedding level, with theory explaining why localized fusion can outperform feature concatenation and prediction averaging. We study this problem in opioid use disorder (OUD), where stress is a major relapse trigger and sustained wearable use can be difficult during early recovery. Using fixed thermal video, FABLE-Therm achieves 0.938 AUROC on held-out participants, and its learned representation transfers to self-reported craving, providing, to our knowledge, the first evidence that craving can be recovered from contactless thermal video. Localized evidence also enables participant-level analysis of deployment failure. We find that improving representation alone is insufficient for equitable deployment: additional data from the underserved group would recover only about half of the cohort gap, while the remainder reflects person-to-person heterogeneity. This modality-agnostic decomposition applies to models with identifiable subpopulations. Together with the first cohort-structured contactless thermal OUD benchmark, our results show that preserving localized evidence supports both accurate sensing and principled analysis of who a model fails and why.

Problem

• Contactless Sensing
• Weakly Supervised Learning
• Evidence Localization
• Opioid Use Disorder
• Fairness

Innovation

• Weakly Supervised Learning
• Thermal Imaging
• Opioid Use Disorder
• Evidence Localization
• Foundation Model Fusion