Encrypted Image Search Based on SGX and Hierarchical Index

Cloud computing meets the storage and computing needs brought about by the explosive growth of images, but it also raises privacy threats. Encryption protects the privacy of outsourced images, but reduces data availability. The paper aims to improve image search accuracy and efficiency while ensuring the security of encrypted images in cloud storage environments. We propose an encrypted image search scheme based on Software Guard Extensions (SGX) and hierarchical index. We design a hierarchical index tree structure to support image search on encrypted domain, which improves search efficiency. We propose an encrypted image search method that combines software and hardware based on SGX. The proposed search method enables the enclave to perform image similarity calculation on plaintext domain, which greatly improves search efficiency and search accuracy. We also delegate the decryption operation of the secret key to SGX, which further achieves one-to-multi access control. The experimental results demonstrate that the proposed scheme greatly improves search efficiency and accuracy.