No abstract

published version features the final layout of the paper including the volume, issue and page numbers. Link to publication General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the "Taverne" license above, please follow below link for the End User

Image Baseline RobustNet [9] ProDA [45] Fully supervised Ground truth WildDash Unseen data IDD KITTI Figure 1. Qualitative results of state-of-the-art domain generalization (RobustNet [9]), unsupervised domain adaptation (ProDA [45]) and fully-supervised [6] methods on unseen datasets WildDash [44], IDD [37], and KITTI [1]. The baseline model and RobustNet are trained only on Cityscapes [10], ProDA is trained on the labeled data of Cityscapes and the unlabeled data of BDD-100K [42] and Mapillary Vistas [25], and the fully-supervised model is trained on the respective 'unseen' dataset. The ProDA method, which leverages unlabeled data, consistently performs best, showing that generalization can be improved by using non-annotated examples.

No abstract