BLOCKCHAIN-BASED PRIVACY-PRESERVING IMAGE DEDUPLICATION AND DECENTRALIZED INTEGRITY AUDITING IN CLOUD STORAGE

Abstract

The rapid growth of digital image collections has increased the storage, communication, and management requirements of cloud storage systems. Conventional data deduplication techniques generally depend on cryptographic hashes, which are effective for detecting identical files but are unable to identify visually similar images having different binary representations. This paper presents a blockchain-based privacy-preserving image deduplication and integrity auditing framework that combines perceptual hashing, secure encryption, ownership verification, cloud storage, and blockchain-based integrity management. The proposed framework preprocesses uploaded images to obtain a standardized representation and generates a perceptual hash (pHash) to capture their visual characteristics. The generated pHash is compared with previously stored hashes using Hamming distance to classify an uploaded image as a new image, exact duplicate, or near-duplicate. For duplicate images, ownership verification is performed before updating the ownership record, thereby avoiding unnecessary physical storage of another copy. New images are encrypted before being stored in the cloud-side database, while associated metadata and ownership information are maintained securely. A SHA-256-based file identifier is generated for reliable file identification, and blockchain smart contracts are used to maintain integrity-related records. The implemented prototype uses Python, SQLite, an Ethereum-compatible blockchain environment, Ganache, Solidity smart contracts, and web3.py. Functional evaluation demonstrates that the framework successfully performs image-aware duplicate detection, encrypted storage, ownership verification, and blockchain-based integrity recording while reducing redundant image storage.