Unsupervised Attention-guided Image-to-Image Translation

Research output: Chapter in Book/Report/Conference proceedingConference contribution

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 that are jointly adversarially trained with the generators and discriminators. We demonstrate qualitatively and quantitatively that our approach attends to relevant regions in the image without requiring supervision, which creates more realistic mappings when compared to those of recent approaches.
LanguageEnglish
Title of host publicationAdvances in Neural Information Processing Systems 31 (NIPS), 2018
Number of pages22
StatusAccepted/In press - 5 Sep 2018

Cite this

Alami Mejjati, Y., Richardt, C., Tompkin, J., Cosker, D., & Kim, K. I. (Accepted/In press). Unsupervised Attention-guided Image-to-Image Translation. In Advances in Neural Information Processing Systems 31 (NIPS), 2018

Unsupervised Attention-guided Image-to-Image Translation. / Alami Mejjati, Youssef; Richardt, Christian; Tompkin, James; Cosker, Darren; Kim, Kwang In.

Advances in Neural Information Processing Systems 31 (NIPS), 2018. 2018.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Alami Mejjati, Y, Richardt, C, Tompkin, J, Cosker, D & Kim, KI 2018, Unsupervised Attention-guided Image-to-Image Translation. in Advances in Neural Information Processing Systems 31 (NIPS), 2018.
Alami Mejjati Y, Richardt C, Tompkin J, Cosker D, Kim KI. Unsupervised Attention-guided Image-to-Image Translation. In Advances in Neural Information Processing Systems 31 (NIPS), 2018. 2018
Alami Mejjati, Youssef ; Richardt, Christian ; Tompkin, James ; Cosker, Darren ; Kim, Kwang In. / Unsupervised Attention-guided Image-to-Image Translation. Advances in Neural Information Processing Systems 31 (NIPS), 2018. 2018.
@inproceedings{4a4207647a1f402c8e428b7e21de2958,
title = "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 that are jointly adversarially trained with the generators and discriminators. We demonstrate qualitatively and quantitatively that our approach attends to relevant regions in the image without requiring supervision, which creates more realistic mappings when compared to those of recent approaches.",
author = "{Alami Mejjati}, Youssef and Christian Richardt and James Tompkin and Darren Cosker and Kim, {Kwang In}",
year = "2018",
month = "9",
day = "5",
language = "English",
booktitle = "Advances in Neural Information Processing Systems 31 (NIPS), 2018",

}

TY - GEN

T1 - Unsupervised Attention-guided Image-to-Image Translation

AU - Alami Mejjati, Youssef

AU - Richardt, Christian

AU - Tompkin, James

AU - Cosker, Darren

AU - Kim, Kwang In

PY - 2018/9/5

Y1 - 2018/9/5

N2 - 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 that are jointly adversarially trained with the generators and discriminators. We demonstrate qualitatively and quantitatively that our approach attends to relevant regions in the image without requiring supervision, which creates more realistic mappings when compared to those of recent approaches.

AB - 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 that are jointly adversarially trained with the generators and discriminators. We demonstrate qualitatively and quantitatively that our approach attends to relevant regions in the image without requiring supervision, which creates more realistic mappings when compared to those of recent approaches.

M3 - Conference contribution

BT - Advances in Neural Information Processing Systems 31 (NIPS), 2018

ER -