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

arXiv cs.CV / 4/16/2026

📰 NewsSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper proposes a physically-guided optical side-channel attack that can exfiltrate electronic screen content without direct contact by projecting and inverting optical signals.
It identifies two key technical instabilities—near-singular projection mappings that break Hadamard stability and light-transport compression that removes global semantic cues—making naive inversion highly sensitive and ambiguous.
The authors introduce IR4Net, which combines a physically regularized irradiance approximation (embedding the radiative transfer equation into a learnable optimizer) with cross-scale contour-to-detail reconstruction to reduce noise propagation.
An irreversibility-constrained semantic reprojection module is used to restore lost global structure by mapping context to semantics.
Experiments across four scene categories show IR4Net achieving higher reconstruction fidelity than competing neural methods and improved robustness to illumination perturbations.

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.