Journal
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021
Volume -, Issue -, Pages 8193-8203Publisher
IEEE COMPUTER SOC
DOI: 10.1109/CVPR46437.2021.00810
Keywords
-
Ask authors/readers for more resources
This paper presents an approach for encoding visual task relationships to improve model performance in an Unsupervised Domain Adaptation (UDA) setting. The study shows that proper encoding of relationships between complementary tasks can improve performance on both tasks. The proposed Cross-Task Relation Layer (CTRL) and Iterative Self-Learning (ISL) training scheme are demonstrated to enhance semantic segmentation and depth estimation tasks in the challenging UDA setting.
We present an approach for encoding visual task relationships to improve model performance in an Unsupervised Domain Adaptation (UDA) setting. Semantic segmentation and monocular depth estimation are shown to be complementary tasks; in a multi-task learning setting, a proper encoding of their relationships can further improve performance on both tasks. Motivated by this observation, we propose a novel Cross-Task Relation Layer (CTRL), which encodes task dependencies between the semantic and depth predictions. To capture the cross-task relationships, we propose a neural network architecture that contains task-specific and cross-task refinement heads. Furthermore, we propose an Iterative Self-Learning (ISL) training scheme, which exploits semantic pseudo-labels to provide extra supervision on the target domain. We experimentally observe improvements in both tasks' performance because the complementary information present in these tasks is better captured. Specifically, we show that: (1) our approach improves performance on all tasks when they are complementary and mutually dependent; (2) the CTRL helps to improve both semantic segmentation and depth estimation tasks performance in the challenging UDA setting; (3) the proposed ISL training scheme further improves the semantic segmentation performance. The implementation is available at https://github.com/susaha/ctrl-uda.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available