Physically-Guided Optical Inversion Enable Non-Contact Side-Channel Attack on Isolated Screens

2026-04-14

AI Summary

This work addresses the challenge of non-intrusive screen content eavesdropping, where reconstruction is typically degraded by the ill-posed instability of projection mapping and irreversible compression in light transport. To overcome these limitations, the authors propose an optical projection side-channel attack paradigm that leverages passive speckle patterns generated via diffuse reflection. The approach integrates a physics-guided neural network (IR4Net) based on the radiative transfer equation, a physically regularized irradiance approximation (PRIrr), and a cross-scale contour-to-detail reconstruction mechanism. Furthermore, an Irreversible Constraint Semantic Reprojection (ICSR) module is introduced to recover global structural coherence. This framework effectively suppresses noise propagation, enhances reconstruction stability, and significantly outperforms existing neural methods across four diverse scenarios, achieving high-fidelity screen content recovery with strong robustness against illumination perturbations.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Non-convex Optimization
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Noncontact exfiltration of electronic screen content poses a security challenge, with side-channel incursions as the principal vector. We introduce an optical projection side-channel paradigm that confronts two core instabilities: (i) the near-singular Jacobian spectrum of projection mapping breaches Hadamard stability, rendering inversion hypersensitive to perturbations; (ii) irreversible compression in light transport obliterates global semantic cues, magnifying reconstruction ambiguity. Exploiting passive speckle patterns formed by diffuse reflection, our Irradiance Robust Radiometric Inversion Network (IR4Net) fuses a Physically Regularized Irradiance Approximation (PRIrr-Approximation), which embeds the radiative transfer equation in a learnable optimizer, with a contour-to-detail cross-scale reconstruction mechanism that arrests noise propagation. Moreover, an Irreversibility Constrained Semantic Reprojection (ICSR) module reinstates lost global structure through context-driven semantic mapping. Evaluated across four scene categories, IR4Net achieves fidelity beyond competing neural approaches while retaining resilience to illumination perturbations.

Problem

• side-channel attack
• optical inversion
• screen content exfiltration
• irreversible compression
• Hadamard stability

Innovation

• Physically-Guided Inversion
• Side-Channel Attack
• Radiative Transfer Equation
• Semantic Reprojection
• Speckle-Based Imaging