Semantically consistent image-to-image translation for unsupervised domain adaptation

  • Unsupervised Domain Adaptation (UDA) aims to adapt models trained on a source domain to a new target domain where no labelled data is available. In this work, we investigate the problem of UDA from a synthetic computer-generated domain to a similar but real-world domain for learning semantic segmentation. We propose a semantically consistent image-to-image translation method in combination with a consistency regularisation method for UDA. We overcome previous limitations on transferring synthetic images to real looking images. We leverage pseudo-labels in order to learn a generative image-to-image translation model that receives additional feedback from semantic labels on both domains. Our method outperforms state-of-the-art methods that combine image-to-image translation and semi-supervised learning on relevant domain adaptation benchmarks, i.e., on GTA5 to Cityscapes and SYNTHIA to Cityscapes.

Download full text files

Export metadata


Number of document requests

Additional Services

Share in Twitter Search Google Scholar
Author:Stephan BrehmORCiDGND, Sebastian SchererGND, Rainer LienhartGND
Frontdoor URL
Parent Title (English):ICAART 2022: Proceedings of the 14th International Conference on Agents and Artificial Intelligence, online streaming, 3-5 February 2022, volume 2
Place of publication:Setúbal
Editor:Ana Paula Rocha, Luc Steels, Jaap van den Herik
Type:Conference Proceeding
Year of first Publication:2022
Publishing Institution:Universität Augsburg
Release Date:2022/02/08
First Page:131
Last Page:141
Institutes:Fakultät für Angewandte Informatik
Fakultät für Angewandte Informatik / Institut für Informatik
Fakultät für Angewandte Informatik / Institut für Informatik / Lehrstuhl für Maschinelles Lernen und Maschinelles Sehen
Dewey Decimal Classification:0 Informatik, Informationswissenschaft, allgemeine Werke / 00 Informatik, Wissen, Systeme / 004 Datenverarbeitung; Informatik
Licence (German):CC-BY-NC-ND 4.0: Creative Commons: Namensnennung - Nicht kommerziell - Keine Bearbeitung (mit Print on Demand)