Detect objects in new domain by distilling more balanced source features with DUA-DA
Detect objects in new domain by distilling more balanced source features with DUA-DA
DUA-DA: Distillation-based Unbiased Alignment for Domain Adaptive Object Detection
arXiv paper abstract https://arxiv.org/abs/2311.10437
arXiv PDF paper https://arxiv.org/pdf/2311.10437.pdf
... Adaptive Object Detection (DAOD) have ... source bias issue, i.e. the aligned features are more favorable towards the source domain, leading to a sub-optimal adaptation.
... domain shift between the source and target domains exacerbates the problem of inconsistent classification and localization in general detection pipelines.
... propose a novel Distillation-based Unbiased Alignment (DUA) framework for DAOD, which can distill the source features towards a more balanced position via a pre-trained teacher model during the training process, alleviating the problem of source bias effectively.
... design a Target-Relevant Object Localization Network (TROLN), which can mine target-related knowledge to produce two classification-free metrics (IoU and centerness).
... implement a Domain-aware Consistency Enhancing (DCE) strategy that utilizes these two metrics to further refine classification confidences, achieving a harmonization between classification and localization in cross-domain scenarios.
... this method ... consistently improves the strong baseline by large margins, outperforming existing alignment-based works.
Please like and share this post if you enjoyed it using the buttons at the bottom!
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
Commentaires