Shallow Features Guide Unsupervised Domain Adaptation for Semantic Segmentation at Class Boundaries

Although deep neural networks have achieved remarkable results for the task of semantic segmentation, they usually fail to generalize towards new domains, especially when performing synthetic-to-real adaptation. Such domain shift is particularly noticeable along class boundaries, invalidating one of...

Full description

Saved in:
Bibliographic Details
Main Authors: Cardace, Adriano, Ramirez, Pierluigi Zama, Salti, Samuele, Di Stefano, Luigi
Format: Conference Proceeding
Language:English
Subjects:
Computer vision
Deep learning
Neural networks
Semantics
Task analysis
Transfer
Few-shot
Semi- and Un- supervised Learning Segmentation
Grouping and Shape
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Although deep neural networks have achieved remarkable results for the task of semantic segmentation, they usually fail to generalize towards new domains, especially when performing synthetic-to-real adaptation. Such domain shift is particularly noticeable along class boundaries, invalidating one of the main goals of semantic segmentation that consists in obtaining sharp segmentation masks.In this work, we specifically address this core problem in the context of Unsupervised Domain Adaptation and present a novel low-level adaptation strategy that allows us to obtain sharp predictions. Moreover, inspired by recent self-training techniques, we introduce an effective data augmentation that alleviates the noise typically present at semantic boundaries when employing pseudo-labels for self-training. Our contributions can be easily integrated into other popular adaptation frameworks, and extensive experiments show that they effectively improve performance along class boundaries.
ISSN:2642-9381
DOI:10.1109/WACV51458.2022.00207