Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation

Unsupervised Domain Adaptation (UDA) serves as a potential alternative for improving cross- domain segmentation tasks. Recent UDA approaches have identified class-wise prototypes and leveraged them to guide the segmentation process in the target domain. However, these methods overlook additional inf...

Full description

Saved in:
Bibliographic Details
Main Authors: Antonio Dauphin Fernando, Thumma Anirudh, Selvaraj Palanisamy, Karthika Prasad, Katia Alexander, Pandiyarasan Veluswamy, Rohini Palanisamy
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11007591/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850132610029715456
author Antonio Dauphin Fernando
Thumma Anirudh
Selvaraj Palanisamy
Karthika Prasad
Katia Alexander
Pandiyarasan Veluswamy
Rohini Palanisamy
author_facet Antonio Dauphin Fernando
Thumma Anirudh
Selvaraj Palanisamy
Karthika Prasad
Katia Alexander
Pandiyarasan Veluswamy
Rohini Palanisamy
author_sort Antonio Dauphin Fernando
collection DOAJ
description Unsupervised Domain Adaptation (UDA) serves as a potential alternative for improving cross- domain segmentation tasks. Recent UDA approaches have identified class-wise prototypes and leveraged them to guide the segmentation process in the target domain. However, these methods overlook additional information from other auxiliary tasks, such as depth, which can potentially improve overall segmentation performance. This paper proposes a Depth-Aware Prototypical learning for Semantic Segmentation (DA- ProSS) pipeline, which includes a novel multitask prototype learning framework that comprises task-specific and task-integrated classifiers. The task-specific classifiers capture the semantic and depth features, and the task-integrated classifier captures the hidden semantic features from depth prediction. This framework ensures that the prototypes from respective heads learn better class-representative features using semantic information and depth cues. Additionally, this pipeline integrates a Self-Refinement learning (SRL) algorithm that generates cross-domain pseudo-labels, which are leveraged to generate refined targets for further self-supervised training. Results indicate that prototypes generated through the depth-encoded semantic task could understand the underlying semantics of the object. The proposed DA-ProSS pipeline with SRL helps the model generalize better to achieve a mIoU of 57.3 that outperforms the previous state-of-the-art methods in the SYNTHIA-to-Cityscapes benchmark dataset.
format Article
id doaj-art-d31b59c27be44520879b70c0164f3f95
institution OA Journals
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-d31b59c27be44520879b70c0164f3f952025-08-20T02:32:10ZengIEEEIEEE Access2169-35362025-01-0113957069571610.1109/ACCESS.2025.357201311007591Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain AdaptationAntonio Dauphin Fernando0Thumma Anirudh1Selvaraj Palanisamy2Karthika Prasad3Katia Alexander4Pandiyarasan Veluswamy5Rohini Palanisamy6https://orcid.org/0000-0001-7050-7681Department of Electronics and Communication Engineering, Indian Institute of Information Technology, Design and Manufacturing, Kancheepuram, Chennai, IndiaDepartment of Electronics and Communication Engineering, Indian Institute of Information Technology, Design and Manufacturing, Kancheepuram, Chennai, IndiaDepartment of Clinical Medicine, Tamil Nadu Veterinary and Animal Sciences University, Chennai, IndiaSchool of Engineering, College of Systems and Society, The Australian National University, Canberra, ACT, AustraliaSchool of Engineering, College of Systems and Society, The Australian National University, Canberra, ACT, AustraliaDepartment of Electronics and Communication Engineering, Indian Institute of Information Technology, Design and Manufacturing, Kancheepuram, Chennai, IndiaDepartment of Electronics and Communication Engineering, Indian Institute of Information Technology, Design and Manufacturing, Kancheepuram, Chennai, IndiaUnsupervised Domain Adaptation (UDA) serves as a potential alternative for improving cross- domain segmentation tasks. Recent UDA approaches have identified class-wise prototypes and leveraged them to guide the segmentation process in the target domain. However, these methods overlook additional information from other auxiliary tasks, such as depth, which can potentially improve overall segmentation performance. This paper proposes a Depth-Aware Prototypical learning for Semantic Segmentation (DA- ProSS) pipeline, which includes a novel multitask prototype learning framework that comprises task-specific and task-integrated classifiers. The task-specific classifiers capture the semantic and depth features, and the task-integrated classifier captures the hidden semantic features from depth prediction. This framework ensures that the prototypes from respective heads learn better class-representative features using semantic information and depth cues. Additionally, this pipeline integrates a Self-Refinement learning (SRL) algorithm that generates cross-domain pseudo-labels, which are leveraged to generate refined targets for further self-supervised training. Results indicate that prototypes generated through the depth-encoded semantic task could understand the underlying semantics of the object. The proposed DA-ProSS pipeline with SRL helps the model generalize better to achieve a mIoU of 57.3 that outperforms the previous state-of-the-art methods in the SYNTHIA-to-Cityscapes benchmark dataset.https://ieeexplore.ieee.org/document/11007591/Contrastive learningprototype learningself-refinement learningunsupervised domain adaptationdepth integration
spellingShingle Antonio Dauphin Fernando
Thumma Anirudh
Selvaraj Palanisamy
Karthika Prasad
Katia Alexander
Pandiyarasan Veluswamy
Rohini Palanisamy
Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation
IEEE Access
Contrastive learning
prototype learning
self-refinement learning
unsupervised domain adaptation
depth integration
title Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation
title_full Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation
title_fullStr Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation
title_full_unstemmed Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation
title_short Depth Integrated Multi-Task Prototypical Learning With Self Refinement for Unsupervised Domain Adaptation
title_sort depth integrated multi task prototypical learning with self refinement for unsupervised domain adaptation
topic Contrastive learning
prototype learning
self-refinement learning
unsupervised domain adaptation
depth integration
url https://ieeexplore.ieee.org/document/11007591/
work_keys_str_mv AT antoniodauphinfernando depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation
AT thummaanirudh depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation
AT selvarajpalanisamy depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation
AT karthikaprasad depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation
AT katiaalexander depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation
AT pandiyarasanveluswamy depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation
AT rohinipalanisamy depthintegratedmultitaskprototypicallearningwithselfrefinementforunsuperviseddomainadaptation