Fuzzy Deduplication: Color-Aware Deduplication for Multi-Media Data

Abstract

Cloud storage technology is constantly evolving, resulting in a significant amount of duplicate data being stored in the cloud, particularly multimedia data such as images and videos. In terms of data privacy and storage optimization, the encrypted deduplication should be checked to save space overhead for cloud servers. Compared to exact deduplication, fuzzy deduplication is low-cost for encrypted multimedia data. Focusing on reducing the false deletion rate, an efficient and secure fuzzy deduplication system based on dual-feature without additional servers is proposed. We also propose a concept of pre-verification for label consistency to compensate for the loss that cannot be fixed through post-verification. Therefore, it is more practical. Finally, we conduct experiments on real-world datasets for performance evaluation. The experimental results show good performance in terms of both computational cost and deduplication efficiency.