Zou Xingxing, Zhang Wen, Zhao Nanxuan
School of Fashion and Textiles, The Hong Kong Polytechnic University, Hong Kong SAR 999077, China.
School of Computing, Informatics, and Decision Systems Engineering, Arizona State University, Tempe, AZ 85281, USA.
J Imaging. 2025 Aug 25;11(9):289. doi: 10.3390/jimaging11090289.
This survey offers a comprehensive overview of advancements in Artificial Intelligence in Graphic Design (AIGD), with a focus on the integration of AI techniques to enhance design interpretation and creative processes. The field is categorized into two primary directions: perception tasks, which involve understanding and analyzing design elements, and generation tasks, which focus on creating new design elements and layouts. The methodology emphasizes the exploration of various subtasks including the perception and generation of visual elements, aesthetic and semantic understanding, and layout analysis and generation. The survey also highlights the role of large language models and multimodal approaches in bridging the gap between localized visual features and global design intent. Despite significant progress, challenges persist in understanding human intent, ensuring interpretability, and maintaining control over multilayered compositions. This survey aims to serve as a guide for researchers, detailing the current state of AIGD and outlining potential future directions.
本次调查全面概述了平面设计中的人工智能(AIGD)进展,重点关注人工智能技术的整合,以增强设计解读和创作过程。该领域分为两个主要方向:感知任务,涉及理解和分析设计元素;生成任务,专注于创建新的设计元素和布局。该方法强调对各种子任务的探索,包括视觉元素的感知和生成、美学和语义理解以及布局分析和生成。该调查还强调了大语言模型和多模态方法在弥合局部视觉特征与全局设计意图之间差距方面的作用。尽管取得了重大进展,但在理解人类意图、确保可解释性以及对多层构图保持控制方面仍存在挑战。本次调查旨在为研究人员提供指导,详细介绍AIGD的当前状态并概述潜在的未来方向。