A nature-inspired edge-cloud collaborative privacy framework for visual relocalization on consumer electronics

Kuang, Yujin, Liu, Jiang, Zou, Minghao, Liu, Yang, Liu, Jing, Yang, Yuan, Zhang, Xiaoguo, Zhou, Wei and Liu, Hantao ORCID: https://orcid.org/0000-0003-4544-3481 2026. A nature-inspired edge-cloud collaborative privacy framework for visual relocalization on consumer electronics. IEEE Transactions on Consumer Electronics 10.1109/TCE.2026.3692722

Full text not available from this repository.

Abstract

To mitigate privacy risks and latency overheads in cloud-based visual relocalization for resource-constrained consumer electronics, this paper proposes an edge-cloud collaborative adaptive privacy-preserving system. We introduce a nature-inspired zero-round-trip-time encryption engine that eliminates the need for network handshakes. Specifically, we design a dual-utility edge hash to modulate the ergodicity of hyper-chaotic systems, which drives a deoxyribonucleic acid dynamic coding mechanism to achieve a robust one-frame-one-key policy. By mapping biomimetic bitwise operations to single instruction multiple data architectures on mobile hardware, we realize high-dimensional feature obfuscation with low latency. To optimize the privacy-utility trade-off, the defense is modeled as a Stackelberg game, where a swarm intelligence algorithm dynamically adjusts protection intensity based on environmental texture entropy. Experimental results on smartphones show that the framework achieves a processing latency of less than 10 ms and a sub-meter localization error of 0.51 meter while effectively thwarting image reconstruction attacks.

Item Type:	Article
Date Type:	Published Online
Status:	In Press
Schools:	Schools > Computer Science & Informatics
Publisher:	Institute of Electrical and Electronics Engineers
ISSN:	0098-3063
Last Modified:	04 Aug 2026 22:04
URI:	https://orca.cardiff.ac.uk/id/eprint/187055

Actions (repository staff only)

Edit Item