생성형 AI 기반 바이오필릭 디자인을 반영한 복합쇼핑몰 라운지 공간 시각화 연구

초록

To address the problem of nature deficit caused by rapid urbanization in modern society, mixed-use retail complexes are evolving into hubs of ‘Retail Therapy’ that provide psychological healing and rest beyond mere consumption spaces. Accordingly, the importance of Biophilic Design, which introduces natural elements into spaces, is emerging; however, there are limitations in expressing its complex organic forms and abstract natural attributes using traditional design tools. The research methods are as follows: First, three spatial expression characteristics capable of visual implementation—‘Physical Form’, ‘Environmental Atmosphere’, and ‘Spatial Experience’—were derived through a literature review. Second, a prompt framework with a ‘Subject + Biophilic Attributes + Parameters’ structure was established to convert these into language recognizable by AI. Third, based on this, visualization experiments were conducted, and an iterative optimization process was performed by analyzing the visual errors of the initial generation and modifying the prompts. As a result of the study, it was confirmed that design visualization using Generative AI secures effectiveness not simply through text-to-image translation, but through an iterative process of prompt control and optimization. First, for ‘Physical Form’, a visualization control method was derived by using architectural integration keywords to restrict the excessive structural complexity arbitrarily generated by AI. Second, for ‘Environmental Atmosphere’, the process of visualizing the intended environment was identified by controlling the disordered spread of natural elements through spatial zoning variables. Third, for ‘Spatial Experience’, a methodology was established to convert complex psychological factors, such as visual protection and openness, into specific visual data through parameter adjustments. This study is significant in that it systematizes a visualization methodology that converts abstract design theories into concrete spatial images, and demonstrates the practical utility of Generative AI as a tool for ideation and communication in the initial design phase.

키워드