Block Chain-Assisted Privacy-Preserving Similarity Image Retrieval Over Cloud

Abstract

With the advancement of digital technology, there has been an exponential increase in the storage of photographs across numerous servers. However, this proliferation of data raises concerns about integrity and reliability, particularly in multi-cloud environments. Existing research on content-based image retrieval has overlooked the challenges posed by untrustworthy cloud servers, hindering effective multi-cloud retrieval. Typically, centralized search servers are introduced in multicloud systems to facilitate image retrieval. Unfortunately, these centralized servers may produce inaccurate or incomplete results, compromising search precision and potentially impacting business interests. Furthermore, cloud servers themselves may exhibit dishonest behavior by tampering with or deleting images to conserve resources, leading to data loss and false image decryption.To address these issues, we propose a solution that ensures both the accuracy of search results and the integrity of images in multi-cloud environments. Our approach involves employing a single-image retrieval technique using blockchain technology to verify both the authenticity of search results and the integrity of stored images. We evaluate the effectiveness of our approach using real-world data, demonstrating its precision and efficiency in combating both over-retrieval and image manipulation. Keywords: Accuracy, Integrity, Privacy Protection, Block Chain, False/Incomplete Results, Multi-Cloud Retrieval, Central Search Server