Title An Analysis of the Applicability of Image-to-3D Generative AI in Architectural Design
Authors 김민석(Kim, Minseok) ; 최현철(Choi, Hyunchul)
DOI https://doi.org/10.5659/JAIK.2026.42.2.209
Page pp.209-216
ISSN 2733-6247
Keywords Generative AI; Image-to-3D; Architectural Design using AI; 3D Shape Reproduction
Abstract This study examines the potential introduction and application of image-to-3D generative AI, which creates three-dimensional models from two-dimensional images, in architectural design. Images of actual buildings, sketches, renderings, and physical model photographs representing regular and irregular single buildings as well as building complexes were input into 12 currently available image-to-3D generative AI tools to produce 3D digital models. The generated outputs were analyzed and compared in terms of accuracy, detail, geometric integrity, and usability. The study then explores possible applications of image-to-3D generative AI in architectural design. The results indicate that these tools demonstrate a strong capacity for 3D shape reproduction and output quality is influenced less by image type or building form and more by the independence, completeness, and ease of inference of the shapes present in the input images.