PPEDCRF: Dynamic-CRF-Guided Selective Perturbation for Background-Based Location Privacy in Video Sequences

2026-04-18

AI Summary

Even when GPS metadata is removed, videos can still be geolocated through background visual cues, posing a significant threat to location privacy. This work proposes a dynamic conditional random field (DCRF)-guided selective perturbation mechanism that identifies sensitive background regions and adaptively adjusts perturbation intensity via normalized control penalty (NCP), injecting Gaussian noise only into critical areas. Evaluated with σ₀=8, the method reduces the Top-1 retrieval accuracy of ResNet18 from 0.667 to 0.361±0.127 while preserving a high visual fidelity of 36.14 dB PSNR—substantially outperforming global noise injection by approximately 6 dB. The approach thus effectively mitigates gallery-based retrieval attacks without compromising perceptual quality.

Technology Category

Computer Vision: Adversarial Attacks & RobustnessMachine Learning: PrivacyKnowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

Security and Privacy: Large-scale security measurementsSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingUser Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

We propose PPEDCRF, a calibrated selective perturbation framework that protects \emph{background-based location privacy} in released video frames against gallery-based retrieval attackers. Even after GPS metadata are stripped, an adversary can geolocate a frame by matching its background visual cues to geo-tagged reference imagery; PPEDCRF mitigates this threat by estimating location-sensitive background regions with a dynamic conditional random field (DCRF), rescaling perturbation strength with a normalized control penalty (NCP), and injecting Gaussian noise only inside the inferred regions via a DP-style calibration rule. On a controlled paired-scene retrieval benchmark with eight attacker backbones and three noise seeds, PPEDCRF reduces ResNet18 Top-1 retrieval accuracy from 0.667 to $0.361\pm0.127$ at $σ_0=8$ while preserving $36.14\,$dB PSNR -- an ${\approx}6\,$dB quality advantage over global Gaussian noise. Transfer across the eight-backbone seed-averaged benchmark is broadly supportive (23 of 24 backbone-gallery cells show negative $Δ$), while appendix-scale confirmation identifies MixVPR as a remaining adverse-transfer exception. Matched-operating-point analysis shows that PPEDCRF and global Gaussian noise converge in Top-1 privacy at equal utility, so the practical benefit is spatially concentrated perturbation that preserves higher visual quality at any given noise scale rather than stronger matched-utility privacy. Code: https://github.com/mabo1215/PPEDCRF

Problem

• location privacy
• background-based privacy
• video sequences
• gallery-based retrieval
• geolocation

Innovation

• Dynamic CRF
• Selective Perturbation
• Location Privacy
• Background-Based Privacy
• DP-Style Calibration