Picking watermarks from noise (PWFN): an improved robust watermarking model against intensive distortions

• DOI: 10.48550/arxiv.2405.05170
• Publication Date: 2024-05-08
• Authors: Sijing Xie; Chengxin Zhao; Nan Sun; Wei Li; Hefei Ling
• Keywords: FOS: Computer and information sciences; Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV); Computer Science - Computer Vision and Pattern Recognition; FOS: Electrical engineering, electronic engineering, information engineering; Electrical Engineering and Systems Science - Image and Video Processing; Computer Science - Multimedia; Multimedia (cs.MM)

Digital watermarking is the process of embedding secret information by altering images in a way that undetectable to human eye. To increase robustness model, many deep learning-based methods use encoder-decoder architecture adding different noises noise layer. The decoder then extracts watermarked from distorted image. However, this method can only resist weak attacks. improve algorithm against stronger noise, paper proposes introduce denoise module between layer and decoder. aimed at reducing recovering some lost during an attack. Additionally, introduces SE fuse pixel-wise channel dimensions-wise, improving encoder's efficiency. Experimental results show our proposed comparable existing models outperforms state-of-the-art under intensities. In addition, ablation experiments superiority module.