Evaluation of Urban Perception Using Only Image Segmentation Features

Xinyi Li, Benjamin Beaucamp, Vincent Tourre, Thomas Leduc, Myriam Servières

2023

Abstract

Deep learning has been used with the street-view imagery Place Pulse 2.0 to evaluate the perception of urban space along six perceptual dimensions: safe, lively, beautiful, wealthy, boring, and depressing. Traditional methods automatically extract feature representations from images through a convolutional neural network to yield prediction. However, the formers are computationally intensive and do not take a priori into account the semantic information from panoptic segmentation scene. In light of this, we propose that learning with semantic information could be close to full image analysis for the prediction of perceptual qualities. A lightweight solution is presented, which quickly predicts the sense of urban space from the implied highly compressed segmentation feature vectors of the street-view images via deep/machine learning models. Our solution achieves an average accuracy of about 62%, which is acceptable compared to the baseline result accuracy of 68%, and significantly reduces the complexity of the data and the computational effort.

Paper Citation

in Harvard Style

Li X., Beaucamp B., Tourre V., Leduc T. and Servières M. (2023). Evaluation of Urban Perception Using Only Image Segmentation Features. In Proceedings of the 9th International Conference on Geographical Information Systems Theory, Applications and Management - Volume 1: GISTAM, ISBN 978-989-758-649-1, SciTePress, pages 200-207. DOI: 10.5220/0011969700003473

in Bibtex Style

@conference{gistam23,
author={Xinyi Li and Benjamin Beaucamp and Vincent Tourre and Thomas Leduc and Myriam Servières},
title={Evaluation of Urban Perception Using Only Image Segmentation Features},
booktitle={Proceedings of the 9th International Conference on Geographical Information Systems Theory, Applications and Management - Volume 1: GISTAM,},
year={2023},
pages={200-207},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011969700003473},
isbn={978-989-758-649-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 9th International Conference on Geographical Information Systems Theory, Applications and Management - Volume 1: GISTAM,
TI - Evaluation of Urban Perception Using Only Image Segmentation Features
SN - 978-989-758-649-1
AU - Li X.
AU - Beaucamp B.
AU - Tourre V.
AU - Leduc T.
AU - Servières M.
PY - 2023
SP - 200
EP - 207
DO - 10.5220/0011969700003473
PB - SciTePress