| Abstrakt: |
House floorplan generation entails crafting efficient spatial layouts within buildings, harmonizing functionality, aesthetics, and usability. The automation of this process is pivotal, expediting design timelines, reducing errors, conserving resources, and facilitating swift exploration of diverse design alternatives for optimal functionality and aesthetics. Nonetheless, the field grapples with inherent challenges, including the provision of diverse layouts to accommodate varied preferences, striking a balance between visual and functional realism, meeting customization demands, and aligning with architectural constraints. In this article, we delve into the transformative impact of computational methods on house floorplan generation. Our study offers a nuanced review and innovative categorization of computational techniques, distinguishing between procedural and deep generative learning approaches. Additionally, we examine representation methods and their interactive capabilities, providing a comprehensive analysis of the advancements, merits, and limitations of contemporary techniques. Furthermore, we critically assess unresolved challenges and delineate promising avenues for future research in computational-based floorplan generation. |
Full text from SpringerLink