Unsupervised Attention-Guided Image-to-Image Translation
Abstract
Current unsupervised image-to-image translation techniques struggle to focus their attention on individual objects without altering the background or the way multiple objects interact within a scene. Motivated by the important role of attention in human perception, we tackle this limitation by introducing unsupervised attention mechanisms which are jointly adversarially trained with the generators and discriminators. We empirically demonstrate that our approach is able to attend to relevant regions in the image without requiring any additional supervision, and that by doing so it achieves more realistic mappings compared to recent approaches.
Cite
Text
Mejjati et al. "Unsupervised Attention-Guided Image-to-Image Translation." Neural Information Processing Systems, 2018.Markdown
[Mejjati et al. "Unsupervised Attention-Guided Image-to-Image Translation." Neural Information Processing Systems, 2018.](https://mlanthology.org/neurips/2018/mejjati2018neurips-unsupervised/)BibTeX
@inproceedings{mejjati2018neurips-unsupervised,
title = {{Unsupervised Attention-Guided Image-to-Image Translation}},
author = {Mejjati, Youssef Alami and Richardt, Christian and Tompkin, James and Cosker, Darren and Kim, Kwang In},
booktitle = {Neural Information Processing Systems},
year = {2018},
pages = {3693-3703},
url = {https://mlanthology.org/neurips/2018/mejjati2018neurips-unsupervised/}
}