Security and privacy protocols for perceptual image hashing
Donghui Hu, Bin Su, Shuli Zheng, Zhong Qiu Zhao, Xintao Wu, Xindong Wu · International Journal of Sensor Networks · 2015
Perceptual hashing plays an increasingly important role in digital forensics, pirate detection and source tracking of the massive networking information. Security and privacy problems exist in existing digital image hashing efforts and have negatively influenced their applications. In this paper we propose and design security and privacy protocols for perceptual image hashing. Specifically we propose the concepts "commutative perceptual hashing and encryption" and "distance preserving feature encryption" to ensure that perceptual hashes in the encryption domain of digital images can be calculated and compared. We further propose three sets of protocols to provide security assurance and privacy protection for the whole process. Each set of protocols are applied on registration, authentication and identification stages of the digital image hashing application models. We improve existing perceptual hashing methods by implementing the security–enhanced protocols. Experimental results demonstrate the effectiveness of the improved perceptual hashing methods.