Verifiable secure image retrieval for cloud-assisted IoT environments

To address the challenges of low accuracy and efficiency in image retrieval, as well as the potential malicious behavior of servers in cloud-assisted Internet of things (IoT) environments, a verifiable and secure image retrieval scheme was proposed. A matrix-based approach was employed for indexing and querying, and the feature matrix was encrypted using an enhanced secure k-nearest neighbor (kNN) algorithm based on the learning with errors (LWE), ensuring improved security in indexing and querying. Blockchain technology was integrated with the quad Merkle Hash tree and efficient short signature to enable the verifiability of search results. Security and performance analyses demonstrated that the security of indexing and querying was ensured by the proposed scheme while the encryption computational complexity and key storage overhead were reduced. The proposed scheme improves image retrieval accuracy and security while optimizing computational and storage resources, making it suitable for cloud-assisted IoT environments.