Quantifying urban streetscapes with deep learning: focus on aesthetic evaluation

TL;DR

This paper presents a deep learning approach to quantify urban streetscape disorder by recognizing facades and billboards, providing a scalable method to assess aesthetic quality based on urban visual elements.

Contribution

The study develops and evaluates a deep learning model for identifying facades and billboards in streetscapes, addressing a gap in scalable aesthetic quantification methods.

Findings

Model achieved 63.17% IoU accuracy in recognizing facades and billboards.

Enables analysis of urban streetscape aesthetics using visual data.

Provides a foundation for integrating aesthetic assessment with urban planning.

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.