Artificial intelligence-designer for high-rise building sketches with user preferences. (15th January 2023)

• Record Type: Journal Article
• Authors: Qian, Wenliang; Yang, Feng; Mei, Hongyuan; Li, Hui
• Is Part Of: Engineering structures. Volume 275(2023)Part A
• Publication Date: 2023-01-15
• Subjects: AI-designer -- High-rise building -- Self-Sparse GAN -- User preference -- Architectural fusion; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2022.115171
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: AI-designer to assist architects for high-rise building sketches. Numerous architectural sketches with a specific theme and diversity generated by AI-designer. User preferences by the geometric shape preferences and the architecture preferences. Abstract: Architectural design is a highly intellectual creative work. It is so interesting and fascinating to explore artificial intelligence (AI)-based architectural design methodology, i.e. exploring the creativity of machines from a design perspective. Conceptual design, which gives a sketch of a building satisfying certain constraints, requires very high creativity and is an important step in the architectural design process. This study proposes a novel AI-based design method of high-rise building sketches with user preferences (termed AI-designer), where AI-designer can control the sketch generation in two aspects: shape preferences and architectural fusion. To achieve these two functions, the AI-based design method contains the shape preference module with Self-Sparse GAN and architecture fusion module. The proposed shape preference module with Self-Sparse GAN enables the user to specify the shape of the building sketch by a simple hand-drawn sketch, where the Self-Similarity score (SS-score) is proposed to obtain the degree of similarity of shapes for selecting a specified number of sketches generated by the Self-Sparse GAN. Meanwhile, the architecture fusion module allows to implement user-preference elements (such as artistic style and local culture) with the generated sketches for further forming a clearer architectural fusion sketch by using the line loss with a feature-based attention mechanism. The design results show that the proposed AI-designer can not only generate a large number of desired architectural sketches with a specific theme while remaining diversity in just a few seconds, but also form clearer sketches with architectural fusion by re-rendering using user-specified elements.