DSD-DA: Distillation-based Source Debiasing for Domain Adaptive Object Detection

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Feng, Yongchao, Li, Shiwei, Gao, Yingjie, Huang, Ziyue, Zhang, Yanan, Liu, Qingjie, Wang, Yunhong
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913353482371072
author Feng, Yongchao
Li, Shiwei
Gao, Yingjie
Huang, Ziyue
Zhang, Yanan
Liu, Qingjie
Wang, Yunhong
author_facet Feng, Yongchao
Li, Shiwei
Gao, Yingjie
Huang, Ziyue
Zhang, Yanan
Liu, Qingjie
Wang, Yunhong
contents Though feature-alignment based Domain Adaptive Object Detection (DAOD) methods have achieved remarkable progress, they ignore the source bias issue, i.e., the detector tends to acquire more source-specific knowledge, impeding its generalization capabilities in the target domain. Furthermore, these methods face a more formidable challenge in achieving consistent classification and localization in the target domain compared to the source domain. To overcome these challenges, we propose a novel Distillation-based Source Debiasing (DSD) framework for DAOD, which can distill domain-agnostic knowledge from a pre-trained teacher model, improving the detector's performance on both domains. In addition, we design a Target-Relevant Object Localization Network (TROLN), which can mine target-related localization information from source and target-style mixed data. Accordingly, we present a Domain-aware Consistency Enhancing (DCE) strategy, in which these information are formulated into a new localization representation to further refine classification scores in the testing stage, achieving a harmonization between classification and localization. Extensive experiments have been conducted to manifest the effectiveness of this method, which consistently improves the strong baseline by large margins, outperforming existing alignment-based works.
format Preprint
id arxiv_https___arxiv_org_abs_2311_10437
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle DSD-DA: Distillation-based Source Debiasing for Domain Adaptive Object Detection
Feng, Yongchao
Li, Shiwei
Gao, Yingjie
Huang, Ziyue
Zhang, Yanan
Liu, Qingjie
Wang, Yunhong
Computer Vision and Pattern Recognition
Though feature-alignment based Domain Adaptive Object Detection (DAOD) methods have achieved remarkable progress, they ignore the source bias issue, i.e., the detector tends to acquire more source-specific knowledge, impeding its generalization capabilities in the target domain. Furthermore, these methods face a more formidable challenge in achieving consistent classification and localization in the target domain compared to the source domain. To overcome these challenges, we propose a novel Distillation-based Source Debiasing (DSD) framework for DAOD, which can distill domain-agnostic knowledge from a pre-trained teacher model, improving the detector's performance on both domains. In addition, we design a Target-Relevant Object Localization Network (TROLN), which can mine target-related localization information from source and target-style mixed data. Accordingly, we present a Domain-aware Consistency Enhancing (DCE) strategy, in which these information are formulated into a new localization representation to further refine classification scores in the testing stage, achieving a harmonization between classification and localization. Extensive experiments have been conducted to manifest the effectiveness of this method, which consistently improves the strong baseline by large margins, outperforming existing alignment-based works.
title DSD-DA: Distillation-based Source Debiasing for Domain Adaptive Object Detection
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2311.10437