Manipulation Localization Using a Hash-Based Proactive Scheme

With the continuous advancement of image tampering techniques, traditional hash-based tampering detection algorithms face numerous challenges, particularly in handling complex tampering types and achieving object-level precise localization. Existing methods often rely on manually designed features, leading to hashes that lack sufficient detail. To address these issues, we propose a hash-based proactive tampering localization approach. First, we design an innovative Hash Construction Module that generates compact yet information-dense image hashes. In the tampering localization module, we combine multi-scale features generated from the tampered image hash with the original and tampered image hashes for precise localization. This approach enhances the interaction and fusion of information, refining the generated features and enabling the detection of even subtle differences in the tampered regions. Experimental results demonstrate that the proposed approach significantly outperforms existing methods in terms of tampering localization accuracy, robustness, and the ability to handle complex tampering types, showcasing strong stability and generalization.