HiFiMSFA: Robust and High-Fidelity Image Watermarking Using Attention Augmented Deep Network

• Published in: IEEE signal processing letters Vol. 32; pp. 781 - 785
• Main Authors: Zhang, Yulin, Ni, Jiangqun, Su, Wenkang
• Format: Journal Article
• Language: English
• Published: New York IEEE 2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: attention mechanism; Convolution; Copy protection; Decoding; deep learning; Digital imaging; Digital rights management; Encoders-Decoders; Feature extraction; Hinge loss; Image quality; Image watermarking; Media; Noise; Robustness; Shape; Social networking (online); Social networks; Training; Watermarking
• ISSN: 1070-9908; 1558-2361
• DOI: 10.1109/LSP.2025.3535216

In recent years, the popularity of digital media sharing, especially high-quality images through online social networks (OSNs) has spurred an increasing demand for digital rights management (DRM) with watermarking. Although the most recent watermarking schemes with deep networks have exhibited considerable performance improvement, they still fall short in resisting multiple attacks with high-fidelity watermarking. To tackle this issue, a customized framework with encoder/decoder structure is proposed in this letter, aiming to consistently improve the robustness performance against multiple attacks. In specific, the M ulti-scale S alient F eature A ttention Block (MSFABlock) is exploited to effectively extract the robust image features with the encoder and decoder by taking advantage of the salient features, e.g., the image features obtained with difference of Gaussian (DoG) and other gradient operators. In addition, an adaptive squared Hinge function is developed as message loss to encourage adaptive watermark embedding. Experimental results demonstrate excellent performance in terms of robustness and perceptual fidelity as well as high efficiency of the proposed scheme in comparison to other SOTA methods.