Ensuring the authenticity and integrity of digital images has become increasingly important due to the widespread availability of advanced image manipulation tools. This paper presents a hybrid framework for image tamper detection and authentication by integrating deep learning-based analysis with blockchain-enabled integrity verification. Error level analysis is employed as a preprocessing step to amplify compression inconsistencies introduced during image manipulation, and the resulting images are classified using a convolutional neural network trained to distinguish between authentic and tampered content. For secure and verifiable authentication, the framework extracts the most significant bit features from image blocks, encrypts and hashes them to construct a Merkle tree, and stores the corresponding Merkle root on a blockchain ledger. The original image data are maintained off-chain using the Inter-Planetary File System to ensure scalability. During verification, recomputed Merkle roots are compared with on-chain records to detect and localize tampered regions. Experimental results on benchmark datasets demonstrate a classification accuracy of 96.21% and effective block-level integrity verification using peak signal-to-noise ratio and mean squared error metrics. The proposed approach provides a reliable and decentralized solution for intelligent image tamper detection and secure authentication.
Guttavelli et al. (2026) studied this question.