Digital Watermarking Technology for AI-Generated Images: A Survey

AI-Generated Content (AIGC) has made significant advancements in both popularity and realism. While the development of generative large models offers immense potential to enhance creativity and operational efficiency, it also introduces a range of risks and challenges, particularly concerning issues such as copyright infringement due to model misuse and the authenticity of generated content. In response to the need for standardized management and application of AIGC models, researchers are increasingly focusing on exploring effective strategies for managing and protecting the authentication of AIGC models, as well as ensuring the traceability of generated images through digital watermarking technologies. This survey provides a comprehensive review of three core areas: the evolution of image generation technologies, traditional and state-of-the-art digital image watermarking algorithms, and watermarking methods specific to AIGC. Additionally, we examine common performance evaluation metrics used in this field. Finally, we discuss the unresolved issues and propose several potential directions for future research.