News
Image integrity control based on digital watermarking, perceptual hashing and digital signature
Abstract
Visual content distributed through media platforms is exposed to the risk of semantic forgery, including manipulation using generative artificial intelligence models. Many proactive protection methods based on digital watermarks are vulnerable to watermark forgery attacks because they lack a cryptographic link between the watermark, the image content, and the holder of the private key. We propose CPW, a method for image integrity verification that integrates a digital signature into the watermark embedding process and links the image content, the embedded authenticator, and a specified public key. A key challenge is the need for bit-exact extraction of long messages under allowable image transformations. To address this, we developed the EPHash perceptual hash function, which provides a robust and compact representation of the image, while the digital signature is protected with an error-correcting code and embedded either as a whole or as independent fragments. Experimental evaluation shows that, compared with MetaSeal, CPW provides higher watermark imperceptibility, while its robustness to allowable transformations depends on the watermark embedding algorithm used. The cryptographic binding of the signature to the EPHash value makes it computationally infeasible to generate a valid authenticator for a different hash value without access to the private key.
Keywords
Edition
Proceedings of the Institute for System Programming, vol. 38, issue 5, 2026, pp. 133-154
ISSN 2220-6426 (Online), ISSN 2079-8156 (Print).
DOI: 10.15514/ISPRAS-2026-38(5)-9
For citation
Full text of the paper in pdf (in Russian)
Back to the contents of the volume