Current Issue Cover
AIGC视觉内容生成与溯源研究进展

刘安安1, 苏育挺1, 王岚君1, 李斌2, 钱振兴3, 张卫明4, 周琳娜5, 张新鹏3, 张勇东4, 黄继武2, 俞能海6(1.天津大学电气自动化与信息工程学院;2.深圳大学电子信息与工程学院;3.复旦大学计算机科学技术学院;4.中国科学技术大学信息科学技术学院;5.北京邮电大学网络空间安全学院;6.中国科学技术大学网络空间安全学院)

摘 要
随着数字媒体与创意产业的快速发展,AIGC技术以其在视觉内容生成中的创新应用而逐渐受到关注。本文旨在围绕AIGC视觉内容生成与溯源研究进展深入研讨。首先,针对图像生成技术进行了探讨,从基于生成式对抗网络的传统方法出发,系统地分析了基于生成式对抗网络、自回归模型和扩散概率模型的最新进展。接着,本文深入探讨可控图像生成技术,突出了通过布局、线稿等附加信息以及基于视觉参考的方法来为创作者提供精确控制的技术现状。随着图像生成技术的革新和应用,生成图像的安全性问题逐渐浮现。而预先审核和过滤的技术手段已难以满足实际需求,故亟需实现生成内容的溯源来进行监管。因此,本文进而对生成图像溯源技术进行研讨,并聚焦水印技术在确保生成内容可靠性和安全性方面的应用。本文依据水印嵌入的流程节点,首先将现有的水印相关的生成图像溯源方法归为无水印嵌入的生成图像溯源、水印前置嵌入的生成图像溯源、水印后置嵌入的生成图像溯源以及联合生成的生成图像溯源并进行详细分析,然后介绍针对生成图像的水印攻击研究现状,最后对生成图像溯源技术进行总结和展望。鉴于视觉内容生成在质量和安全上的挑战,本文旨在为研究者提供一个视觉内容生成与溯源的系统研究视角,以促进数字媒体创作环境的安全与可信,并引导未来相关技术的发展方向。
关键词
Review on the progress of the AIGC visual content generation and traceability

Liu Anan, Su Yuting1, Wang Lanjun1, Li Bin2, Qian Zhenxing3, Zhang Weiming4, Zhou Linna5, Zhang Xinpeng3, Zhang Yongdong4, Huang Jiwu2, Yu Nenghai6(1.School of Electrical and Information Engineering,Tianjin University;2.College of Electronics and Information Engineering,Shenzhen University;3.School of Computer Science,Fudan University;4.School of Information Science and Technology,University of Science and Technology of China;5.School of Cyberspace Security,Beijing University of Posts and Telecommunications;6.School of Cyber Science and Technology,University of Science and Technology of China)

Abstract
In the contemporary digital era, characterized by rapid technological advancements, multimedia content creation, particularly in visual content generation, has become an integral part of modern societal development. The exponential growth of digital media and the creative industry has thrust Artificial Intelligence Generated Content (AIGC) technology into the limelight. AIGC"s groundbreaking applications in visual content generation have not only equipped multimedia creators with novel tools and capabilities but have also delivered substantial benefits across diverse domains, spanning from the realms of cinema and gaming to the immersive landscapes of virtual reality. This review gives on a comprehensive introduction to delve into the profound advancements within Artificial Intelligence Generated Content (AIGC) technology. Our particular emphasis is on the domain of visual content generation and its critical facet of traceability. Initially, our discussions traces the evolutionary path of image generation technology, from its inception within Generative Adversarial Networks (GANs) to the latest advancements in Transformer auto-regressive models and diffusion probability models. This progression unveils a remarkable leap in both the quality and capability of image generation, underscoring the rapid evolution of this field, which has transitioned from its nascent stages to an era characterized by explosive growth. In this section, we delve into the development of GANs, encompassing their evolution from text-conditioned methods to sophisticated techniques for style control and the development of large-scale models. Furthermore, we explore the emergence of Transformer-based auto-regressive models, exemplified by the likes of DALL-E and CogView, which have heralded a new epoch in the domain of image generation. Additionally, our discourse delves into the burgeoning interest surrounding diffusion probability models, renowned for their stable training methods and their ability to yield high-quality outputs. As the development of Artificial Intelligence Generated Content (AIGC) technology continues to advance, it encounters challenges, prominently among them being the enhancement of content quality and the imperative of precise control to align with specific requisites. Within this context, this review conducts a thorough exploration of controllable image generation technology, a pivotal research domain that strives to furnish meticulous control over the generated content. This achievement is facilitated through the integration of supplementary elements, such as intricate layouts, detailed sketches, and precise visual references, thereby empowering creators to preserve their artistic autonomy while upholding exacting standards of quality. One notable facet that has garnered considerable academic attention is the utilization of visual references as a mechanism to enable the generation of diverse styles and personalized outcomes by incorporating user-provided visual elements. This review underscores the profound potential inherent in these methodologies, illustrating their transformative role across domains such as digital art and interactive media. Within the development of these technologies, while ushering in new horizons in digital creativity, simultaneously presents profound challenges, particularly in the domain of image authenticity and the potential for malevolent misuse, exemplified by the creation of deepfakes or the proliferation of fake news. These challenges extend far beyond mere technical intricacies; they encompass substantial risks pertaining to individual privacy, security, and the broader societal implications of eroding public trust and social stability. In response to these formidable challenges, watermark-related image traceability technology has emerged as an indispensable solution. This technology harnesses the power of watermarking techniques to authenticate and verify AI-generated images, thereby safeguarding their integrity. Within the pages of this review, we meticulously categorize these watermarking techniques into distinct types, encompassing watermark-free embedding, watermark pre-embedding, watermark post-embedding, and joint generation methods. Each of these approaches plays a pivotal role in the verification of traceability across diverse scenarios, thereby offering a robust defense against potential misuses of AI-generated imagery. In conclusion, while AIGC technology offers promising new opportunities in visual content creation, it simultaneously brings to the fore significant challenges regarding the security and integrity of generated content. This comprehensive review covers the breadth of AIGC technology, starting from an overview of existing image generation technologies, such as GANs, auto-regressive models, and diffusion probability models. It then categorizes and analyzes controllable image generation technology from the perspectives of additional conditions and visual examples. Additionally, the review focuses on watermark-related image traceability technology, discussing various watermark embedding techniques, the current state of watermark attacks on generated images, and providing an extensive overview and future outlook of generation image traceability technology. The aim is to offer researchers a detailed, systematic, and comprehensive perspective on the advancements in AIGC visual content generation and traceability, deepening the understanding of current research trends, challenges, and future directions in this rapidly evolving field.
Keywords

订阅号|日报