Cross-Domain Autonomous Driving Perception using Contrastive Appearance Adaptation

Ziqiang Zheng, Yingshu Chen, Binh-Son Hua, Yang Wu, Sai-Kit Yeung

Keywords: Domain adaptation, Object detection, Semantic segmentation, Depth estimation

Abstract: Addressing domain shifts for complex perception tasks in autonomous driving has long been a challenging problem. In this paper, we show that existing domain adaptation methods pay little attention to the \textit{content mismatch} issue between source and target images, thereby weakening the domain adaptation performance and the decoupling of domain-invariant and domain-specific representations. To solve the aforementioned problems, we propose an image-level domain adaptation framework that aims at adapting source-domain images to the target domain with content-aligned image pairs. Our framework consists of three mutual-beneficial modules in a cycle: a \textit{cross-domain content alignment} module to generate source-target pairs with consistent content representations in a self-supervised manner, \textit{a reference-guided image synthesis} using the generated content-aligned source-target image pairs, and a \textit{contrastive learning} module to self-supervise domain-invariant feature extractor from the generated images. Our contrastive appearance adaptation is task-agnostic and robust to complex perception tasks in autonomous driving. Our proposed method demonstrates state-of-the-art results in cross-domain object detection, semantic segmentation, and depth estimation as well as better image synthesis ability qualitatively and quantitatively.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Applications (eg, speech processing, computer vision, NLP)