Pixel and Feature Level Based Domain Adaption for Object Detection in Autonomous Driving

TL;DR

This paper introduces a novel unsupervised domain adaptation approach for object detection in autonomous driving, combining pixel and feature level transformations to improve generalization from synthetic to real datasets.

Contribution

It proposes an integrated model using pixel and feature level adaptations with end-to-end training, addressing semantic consistency and domain shift in complex object detection tasks.

Findings

Outperforms existing methods on multiple datasets

Demonstrates robustness in cross-domain object detection

Effectively preserves semantic information during adaptation

Abstract

Annotating large scale datasets to train modern convolutional neural networks is prohibitively expensive and time-consuming for many real tasks. One alternative is to train the model on labeled synthetic datasets and apply it in the real scenes. However, this straightforward method often fails to generalize well mainly due to the domain bias between the synthetic and real datasets. Many unsupervised domain adaptation (UDA) methods are introduced to address this problem but most of them only focus on the simple classification task. In this paper, we present a novel UDA model to solve the more complex object detection problem in the context of autonomous driving. Our model integrates both pixel level and feature level based transformtions to fulfill the cross domain detection task and can be further trained end-to-end to pursue better performance. We employ objectives of the generative adversarial network and the cycle consistency loss for image translation in the pixel space. To address the potential semantic inconsistency problem, we propose region proposal based feature adversarial training to preserve the semantics of our target objects as well as further minimize the domain shifts. Extensive experiments are conducted on several different datasets, and the results demonstrate the robustness and superiority of our method.