Unsupervised Domain Adaptation On Cityscapes 1

Métriques

Résultats

Résultats de performance de divers modèles sur ce benchmark

Nom du modèle	[email protected]	Paper Title	Repository
SCDA	33.8	Adapting Object Detectors via Selective Cross-Domain Alignment
SAPNet	40.9	Spatial Attention Pyramid Network for Unsupervised Domain Adaptation	-
UMT (ResNet50-FPN, 1024px)	61.4	Align and Distill: Unifying and Improving Domain Adaptive Object Detection	-
DDMRL	34.6	Diversify and Match: A Domain Adaptive Representation Learning Paradigm for Object Detection	-
DDF	42.3	Decompose to Adapt: Cross-domain Object Detection via Feature Disentanglement	-
ViSGA	43.3	Seeking Similarities over Differences: Similarity-based Domain Alignment for Adaptive Object Detection	-
PT (ResNet50-FPN, 1024px)	59.2	Align and Distill: Unifying and Improving Domain Adaptive Object Detection	-
AWADA	44.8	AWADA: Attention-Weighted Adversarial Domain Adaptation for Object Detection	-
MCAR	38.8	Adaptive Object Detection with Dual Multi-Label Prediction	-
O2net	46.8	Improving Transferability for Domain Adaptive Detection Transformers	-
DA-RetinaNet	41.87	An Unsupervised Domain Adaptation Scheme for Single-Stage Artwork Recognition in Cultural Sites	-
SWDA	34.8	Strong-Weak Distribution Alignment for Adaptive Object Detection	-
SADA (ResNet50-FPN, 1024px)	54.2	Align and Distill: Unifying and Improving Domain Adaptive Object Detection	-
ALDI-DETR (ResNet-50, 800px)	44.8	Align and Distill: Unifying and Improving Domain Adaptive Object Detection	-
ILLUME	43.8	To miss-attend is to misalign! Residual Self-Attentive Feature Alignment for Adapting Object Detectors
ALDI-YOLO (1024px)	62.5	Align and Distill: Unifying and Improving Domain Adaptive Object Detection	-
SAD	45.2	Self-Adversarial Disentangling for Specific Domain Adaptation	-
LGCL (unsupervised)	45.3	Improving Object Detection via Local-global Contrastive Learning	-
AT (ResNet50-FPN, 1024px)	63.3	Align and Distill: Unifying and Improving Domain Adaptive Object Detection	-
SSA-DA	42.5	Bi-Dimensional Feature Alignment for Cross-Domain Object Detection	-

0 of 26 row(s) selected.