A study on the automatic generation of banner layouts. (July 2021)

• Record Type: Journal Article
• Authors: Hu, Hao; Zhang, Chao; Liang, Yanxue
• Is Part Of: Computers & electrical engineering. Volume 93(2021)
• Publication Date: 2021-07
• Subjects: banner design -- layout generation -- traditional method -- deep reinforcement learning; Computer engineering -- Periodicals; Electrical engineering -- Periodicals; Electrical engineering -- Data processing -- Periodicals; Ordinateurs -- Conception et construction -- Périodiques; Électrotechnique -- Périodiques; Électrotechnique -- Informatique -- Périodiques; Computer engineering; Electrical engineering; Electrical engineering -- Data processing; Periodicals; Electronic journals; 621.302854
• DOI: 10.1016/j.compeleceng.2021.107269
• Languages: English
• ISSNs: 0045-7906

Abstract

Highlights: Filling design elements to the layout to generate a banner Aesthetic measures to evaluate the design layout Traditional method to optimize the design layout based on designers' experience Optimize the design layout automatically with deep reinforcement learning Abstract: This paper addresses the automatic banner design problem as a layout generation task, where a banner is generated by filling design elements to a given layout. We initially introduce the traditional method of generating design layouts used by designers. After that, we introduce a deep reinforcement learning (DRL) based method to learn the policy of generating design layouts. Evaluation metrics are introduced to assess the quality of design layouts generated by the two proposed methods. The experiment shows that, both methods can generate a layout for a banner of given size, and the DRL based method outperforms the traditional method by generating layouts of better quality. However, results based on the DRL method are affected by reward definitions and a long training process is needed to achieve a better performance. Graphical abstract: Image, graphical abstract