Quantifying urban streetscapes with deep learning: focus on aesthetic evaluation

2021/06/29 by Yusuke Kumakoshi, Kumakoshi, Yusuke, Shigeaki Onoda +5
Engineering · Environmental Science · #Computer Vision and Pattern Recognition (cs.CV) #Computers and Society (cs.CY) #FOS: Computer and information sciences #Land Use and Ecosystem Services #Urban Design and Spatial Analysis #Urban Green Space and Health

paper · pdf · doi:10.48550/arxiv.2106.15361

openalex publication_date 2021/06/29 · openalex created_date 2021/07/05 · openalex updated_date 2026/07/28

Abstract

The disorder of urban streetscapes would negatively affect people's perception of their aesthetic quality. The presence of billboards on building facades has been regarded as an important factor of the disorder, but its quantification methodology has not yet been developed in a scalable manner. To fill the gap, this paper reports the performance of our deep learning model on a unique data set prepared in Tokyo to recognize the areas covered by facades and billboards in streetscapes, respectively. The model achieved 63.17 % of accuracy, measured by Intersection-over-Union (IoU), thus enabling researchers and practitioners to obtain insights on urban streetscape design by combining data of people's preferences.

Citations

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Quantifying Urban Canopy Cover with Deep Convolutional Neural Networks
• Hierarchical Multi-Scale Attention for Semantic Segmentation

Related

• A machine learning method for the large-scale evaluation of urban visual environment
• Deep Learning the City : Quantifying Urban Perception At A Global Scale
• Streetscape Analysis with Generative AI (SAGAI): Vision-Language Assessment and Mapping of Urban Scenes
• Deep learning reveals cross-platform consistency in street view imagery for urban perception mapping
• Interpretable Multimodal Framework for Human-Centered Street Assessment: Integrating Visual-Language Models for Perceptual Urban Diagnostics