Image-to-Image Translation for Autonomous Driving from Coarsely-Aligned Image Pairs
This addresses the challenge of reliable autonomous driving in adverse weather conditions, but it is incremental as it builds on prior unpaired image-to-image translation methods by leveraging coarse alignments.
The paper tackles the problem of improving autonomous driving performance in adverse weather by translating sensor images from adverse to benign conditions, using coarsely-aligned image pairs, and shows that this approach enhances image translation quality and boosts accuracy in downstream tasks like semantic segmentation.
A self-driving car must be able to reliably handle adverse weather conditions (e.g., snowy) to operate safely. In this paper, we investigate the idea of turning sensor inputs (i.e., images) captured in an adverse condition into a benign one (i.e., sunny), upon which the downstream tasks (e.g., semantic segmentation) can attain high accuracy. Prior work primarily formulates this as an unpaired image-to-image translation problem due to the lack of paired images captured under the exact same camera poses and semantic layouts. While perfectly-aligned images are not available, one can easily obtain coarsely-paired images. For instance, many people drive the same routes daily in both good and adverse weather; thus, images captured at close-by GPS locations can form a pair. Though data from repeated traversals are unlikely to capture the same foreground objects, we posit that they provide rich contextual information to supervise the image translation model. To this end, we propose a novel training objective leveraging coarsely-aligned image pairs. We show that our coarsely-aligned training scheme leads to a better image translation quality and improved downstream tasks, such as semantic segmentation, monocular depth estimation, and visual localization.