… To solve this concern, we propose our novel zigzag-learn and incor-know components. We also conduct an ablation test, comparing ours with AT [4] in RGB→thermal DA test …
D Phat Do, T Kim, J Na, J Kim, K Lee, K Cho… - arXiv e …, 2024 - ui.adsabs.harvard.edu
… propose a DistinctiveDual-DomainTeacher (D3T) framework that employs distinct training … Specifically, we segregate the source and target training sets for building dual-teachers and …
Z Chen, M Hu, B Chen - Available at SSRN 4924627 - papers.ssrn.com
… moments, the RGB and thermal images present distinct appearances, as shown in Fig. 5. … D3t: Distinctive dual-domainteacherzigzaggingacrossrgb-thermalgap for domain-adaptive …
Y Cui, L Li, J Zhang, C Yan, H Wang, S Wang… - Proceedings of the …, 2024 - dl.acm.org
… teacher model through EMA, thereby improving the quality of pseudo-labels and creating a virtuous cycle. Concretely, the teacher … with a large domain gap. SOCCER achieves 63.8% AP…
P Gurunath, S Udupa, A Gandhamal, S Madhu… - arXiv preprint arXiv …, 2024 - arxiv.org
… Contrastive mean teacher for domainadaptive object detectors. In Proceedings of the IEEE/… D3t: Distinctivedual-domainteacherzigzaggingacrossrgb-thermalgap for domain-adaptive …
… D3T: DistinctiveDual-DomainTeacherZigzaggingAcrossRGB-ThermalGap for Domain-Adaptive Object Detection. In Proceedings of the IEEE/CVF Conference on Computer Vision …