A Novel Unsupervised Domain Adaption Method for Depth-Guided Semantic Segmentation Using Coarse-to-Fine Alignment

Domain adaptation methods in machine learning deal with the domain shift issue by aligning source and target data representation. This paper proposes a novel domain adaptation method for semantic segmentation that exploits the Fourier transform on chromatic space to improve the quality of style tran...

Full description

Saved in:
Bibliographic Details
Published in:IEEE access 2022, Vol.10, p.101248-101262
Main Authors: Nam, Kieu Dang, Nguyen, Tu M., Dieu, Trinh V., Visani, Muriel, Nguyen, Thi-Oanh, Sang, Dinh Viet
Format: Article
Language:English
Subjects:
Adaptation
Adaptation models
adversarial learning
Alignment
Data models
Deep learning
domain adaptation
Domains
Estimation
Fourier transforms
Image segmentation
Machine learning
Semantic segmentation
Semantics
style transfer
Task analysis
Training
Training data
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Domain adaptation methods in machine learning deal with the domain shift issue by aligning source and target data representation. This paper proposes a novel domain adaptation method for semantic segmentation that exploits the Fourier transform on chromatic space to improve the quality of style transfer, and generates pseudo-labels for self-training by combining the results from different teachers obtained at different rounds of self-training. Our method also applies class-level adversarial learning to achieve a more fine-grained alignment between the two domains, and a late fusion with a depth-estimation model to improve its segmentation outputs. Experiments show that our method yields superior performance in terms of accuracy compared to other existing state-of-the-art methods.
ISSN:2169-3536
2169-3536
DOI:10.1109/ACCESS.2022.3205414