Learning a Domain-Agnostic Visual Representation for Autonomous Driving via Contrastive Loss
This addresses the costly data annotation issue for autonomous driving by improving generalization from synthetic to real-world data, though it is incremental as it builds on existing domain adaptation techniques.
The paper tackles the domain shift problem in autonomous driving by proposing a two-stage unsupervised domain adaptation framework with cyclic adversarial training and contrastive loss, achieving better performance in monocular depth estimation compared to previous state-of-the-art methods and showing effectiveness in semantic segmentation.
Deep neural networks have been widely studied in autonomous driving applications such as semantic segmentation or depth estimation. However, training a neural network in a supervised manner requires a large amount of annotated labels which are expensive and time-consuming to collect. Recent studies leverage synthetic data collected from a virtual environment which are much easier to acquire and more accurate compared to data from the real world, but they usually suffer from poor generalization due to the inherent domain shift problem. In this paper, we propose a Domain-Agnostic Contrastive Learning (DACL) which is a two-stage unsupervised domain adaptation framework with cyclic adversarial training and contrastive loss. DACL leads the neural network to learn domain-agnostic representation to overcome performance degradation when there exists a difference between training and test data distribution. Our proposed approach achieves better performance in the monocular depth estimation task compared to previous state-of-the-art methods and also shows effectiveness in the semantic segmentation task.