| Title |
Analyzing Spatial Programs and Design Elements Using Deep Learning |
| Authors |
한유진(Han, Yoojin) ; 이현수(Lee, Hyunsoo) |
| DOI |
https://doi.org/10.5659/JAIK.2023.39.9.105 |
| Keywords |
Deep Learning; Computer Vision; CNN; Image Classification; Social Big Data; Spatial Design; Hotel Design |
| Abstract |
This study aims to uncover the essential spatial programs and design elements that resonate with lifestyle hotel users. It utilizes deep learning
methods with social big data to access authentic customer opinions in today’s digital world. In this context, this research focuses on
evaluating Instagram images of South Korean lifestyle hotels systematically collected using a Python web crawler developed by the researcher.
The image dataset was initially analyzed using a pre-built computer vision model to explore spatial design elements. Subsequently,
Convolutional Neural Networks (CNN) was applied to scrutinize images categorized as spatial in the previous stage, identifying crucial spatial
programs. These findings emphasize the significance of decorative elements like furnishings, materials, textiles, and indoor greenery in shaping
lifestyle hotel environments. Additionally, this research revealed that these hotels offer a range of services beyond accommodation, with a
strong emphasis on Food and Beverage (F&B), banqueting facilities, and retail offerings. Ultimately, this study aims to expand and enrich the
toolbox of big data analysis techniques and deep learning models in the field of architecture and spatial design, introducing a new paradigm
for their application. |