Communications of the ACM
Social Media, AI Can Measure Aesthetic Quality of Landscapes
To develop their model, the scientists trained a deep-learning algorithm on over 200,000 photos of landscapes in Great Britain obtained from the Scenic-Or-Not database.
Credit: iStock
Researchers at Switzerland's École polytechnique fédérale de Lausanne (EPFL) and the Netherlands' Wageningen University used deep learning to develop a model that can factor people's aesthetic enjoyment of a landscape into ecosystem service assessments.
The researchers trained the algorithm on more than 200,000 photos of U.K. landscapes from the Scenic-Or-Not database, which were rated via crowdsourcing as to their aesthetic quality.
The model then was run on more than 9 million photos of U.K. landscapes posted on Flickr to predict scenic beauty.
The researchers found that at a resolution of 500 square meters, the Flickr model is more accurate than the conventional environmental indicator-based model.
EPFL's Devis Tuia said, "This indicates that the use of social media-based data provides a combination of information about the state of the environment and how people interact with it. Such information has never before been obtained with such a high degree of accuracy."
From EPFL News (Switzerland)
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found