Aesthetic Features for Personalized Photo Recommendation
Abstract
Many photography websites such as Flickr, 500px, Unsplash, and Adobe Behance are used by amateur and professional photography enthusiasts. Unlike content-based image search, such users of photography websites are not just looking for photos with certain content, but more generally for photos with a certain photographic "aesthetic". In this context, we explore personalized photo recommendation and propose two aesthetic feature extraction methods based on (i) color space and (ii) deep style transfer embeddings. Using a dataset from 500px, we evaluate how these features can be best leveraged by collaborative filtering methods and show that (ii) provides a significant boost in photo recommendation performance.
Cite
@article{arxiv.1809.00060,
title = {Aesthetic Features for Personalized Photo Recommendation},
author = {Yu Qing Zhou and Ga Wu and Scott Sanner and Putra Manggala},
journal= {arXiv preprint arXiv:1809.00060},
year = {2018}
}
Comments
In Proceedings of the Late-Breaking Results track part of the Twelfth ACM Conference on Recommender Systems, Vancouver, BC, Canada, October 6, 2018, 2 pages