An Effective Anti-Object-Detection Image Privacy Protection Scheme Based on Robust Chaos
Yuexi Peng, Zixin Lan, Zhijun Li, Chunlai Li
Abstract
Incidents of cloud data leakage by internet companies have occurred frequently, and the issue of image privacy has gradually attracted people's attention. Traditional image encryption protects the privacy of image while sacrificing the visual usability, which is obviously not suitable for cloud storage. A novel encryption scheme combining traditional encryption with thumbnail-preserving encryption is proposed. The proposed encryption scheme divides the original image into visual and nonvisual images, and uses different schemes for encryption processing. Users can choose through visual images at cloud service, and nonvisual images are hidden in invisible databases. A series of representative thumbnail-preserving encryption experiments have proven the superiority of the scheme. Furthermore, object detection has been used for the first time. Compared with existing methods, the proposed scheme also exhibits superior performance in the same environment.