Presentation Information
[C4-01]Evaluating the gender and age differences in the subjective perceptions of streetscapes
*Dongou Yang1,2, Chenbo Zhao3,1, Yoshiki Ogawa1, Takuya Oki4, Yoshihide Sekimoto1 (1. Center for Spatial Information Science (CSIS), The University of Tokyo, 2. Graduate School of Frontier Sciences, The University of Tokyo, 3. Department of Civil Engineering, The University of Tokyo, 4. School of Environment and Society, Tokyo Institute of Technology)
Keywords:
subjective perceptions,streetscapes,CLIP,gender and age differences
This study explores the variations in subjective perceptions of streetscapes based on gender and age. Evaluating people’s subjective perceptions of streetscapes is significant for urban planning. However, there are few previous studies that conduct a quantitative analysis on the gender and age differences in various kinds of subjective perceptions. Furthermore, deep learning models used in previous studies are not specialized for studying the relationship between text and images. In this article, we use a neural network named CLIP, which is efficient and performs well in tasks involving the relationship between text and images. Our model can maintain low computational cost while achieving high accuracy in measuring subjective perceptions on streetscapes from different gender and age groups. In our results, we reveal the differences in how male and female, as well as different age groups, experience and prioritize elements of urban streetscapes.