Saved in:
Bibliographic Details
Main Authors: Zeng, Ziru, Ding, Yue, Lu, Hongtao
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2401.13264
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910306583707648
author Zeng, Ziru
Ding, Yue
Lu, Hongtao
author_facet Zeng, Ziru
Ding, Yue
Lu, Hongtao
contents Recently,the detection transformer has gained substantial attention for its inherent minimal post-processing requirement.However,this paradigm relies on abundant training data,yet in the context of the cross-domain adaptation,insufficient labels in the target domain exacerbate issues of class imbalance and model performance degradation.To address these challenges, we propose a novel class-aware cross domain detection transformer based on the adversarial learning and mean-teacher framework.First,considering the inconsistencies between the classification and regression tasks,we introduce an IoU-aware prediction branch and exploit the consistency of classification and location scores to filter and reweight pseudo labels.Second, we devise a dynamic category threshold refinement to adaptively manage model confidence.Third,to alleviate the class imbalance,an instance-level class-aware contrastive learning module is presented to encourage the generation of discriminative features for each class,particularly benefiting minority classes.Experimental results across diverse domain-adaptive scenarios validate our method's effectiveness in improving performance and alleviating class imbalance issues,which outperforms the state-of-the-art transformer based methods.
format Preprint
id arxiv_https___arxiv_org_abs_2401_13264
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing cross-domain detection: adaptive class-aware contrastive transformer
Zeng, Ziru
Ding, Yue
Lu, Hongtao
Computer Vision and Pattern Recognition
Recently,the detection transformer has gained substantial attention for its inherent minimal post-processing requirement.However,this paradigm relies on abundant training data,yet in the context of the cross-domain adaptation,insufficient labels in the target domain exacerbate issues of class imbalance and model performance degradation.To address these challenges, we propose a novel class-aware cross domain detection transformer based on the adversarial learning and mean-teacher framework.First,considering the inconsistencies between the classification and regression tasks,we introduce an IoU-aware prediction branch and exploit the consistency of classification and location scores to filter and reweight pseudo labels.Second, we devise a dynamic category threshold refinement to adaptively manage model confidence.Third,to alleviate the class imbalance,an instance-level class-aware contrastive learning module is presented to encourage the generation of discriminative features for each class,particularly benefiting minority classes.Experimental results across diverse domain-adaptive scenarios validate our method's effectiveness in improving performance and alleviating class imbalance issues,which outperforms the state-of-the-art transformer based methods.
title Enhancing cross-domain detection: adaptive class-aware contrastive transformer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2401.13264