Ciphertext Fuzzy Retrieval Mechanism With Bidirectional Verification and Privacy Protection

Searchable encryption (SE) and attribute-based SE (ABSE) enable accurate encrypted data searches in cloud storage but face challenges, such as supporting only unidirectional verification. Additionally, ABSE lacks support for fuzzy search and unduly restricts search privileges, while SE schemes suffer from low-search efficiency and poor fuzzy search accuracy. To address these issues, the proposed scheme introduces a fuzzy retrieval mechanism with bidirectional verification and privacy protection. We constructed an on-chain and off-chain collaborative storage model to ensure the secure storage of massive data while enhancing access control flexibility by separating search authority from decryption authority. Additionally, we proposed a high-precision multikeyword fuzzy search scheme that integrates an improved secure k-nearest neighbor (SKNN) algorithm, the SimHash algorithm, and the random projection locality-sensitive hash (RPLSH) model. We also designed a bidirectional verification strategy resistant to collusion attacks to prevent data owner repudiation and internal collusion. Finally, we proposed a personalized extreme privacy protection strategy to ensure differential privacy security for search results. Security analysis shows that the proposed scheme satisfies adaptive chosen-keyword semantic security and resists collusion attacks, effectively protecting user privacy and data security. Experimental results demonstrate that the precision rate of the proposed scheme improves by 14.4% and 9.3% compared to the traditional LSH+BF fuzzy search architecture and the VRFMS scheme, respectively. Additionally, the search efficiency of the proposed scheme increases by approximately 36% and 54% compared to the MKFS and FWSR schemes, respectively.