HyperAIHyperAI超神经
首页算力平台文档资讯论文教程数据集百科SOTALLM 模型天梯GPU 天梯顶会
全站搜索
关于
HyperAIHyperAI超神经

Unsupervised Domain Adaptation On Cityscapes 1

评估指标

mAP@0.5

评测结果

各个模型在此基准测试上的表现结果

Paper TitleRepository
ALDI++ (ResNet50-FPN, 1024px)66.8Align and Distill: Unifying and Improving Domain Adaptive Object Detection
AT (ResNet50-FPN, 1024px)63.3Align and Distill: Unifying and Improving Domain Adaptive Object Detection
ALDI-YOLO (1024px)62.5Align and Distill: Unifying and Improving Domain Adaptive Object Detection
MIC (ResNet50-FPN, 1024px)61.7Align and Distill: Unifying and Improving Domain Adaptive Object Detection
UMT (ResNet50-FPN, 1024px)61.4Align and Distill: Unifying and Improving Domain Adaptive Object Detection
PT (ResNet50-FPN, 1024px)59.2Align and Distill: Unifying and Improving Domain Adaptive Object Detection
SADA (ResNet50-FPN, 1024px)54.2Align and Distill: Unifying and Improving Domain Adaptive Object Detection
MRT51.2Masked Retraining Teacher-Student Framework for Domain Adaptive Object Detection-
MIC47.6MIC: Masked Image Consistency for Context-Enhanced Domain Adaptation
O2net46.8Improving Transferability for Domain Adaptive Detection Transformers
LGCL (supervised)46.7Improving Object Detection via Local-global Contrastive Learning-
LGCL (unsupervised)45.3Improving Object Detection via Local-global Contrastive Learning-
SAD45.2Self-Adversarial Disentangling for Specific Domain Adaptation-
AWADA44.8AWADA: Attention-Weighted Adversarial Domain Adaptation for Object Detection-
ALDI-DETR (ResNet-50, 800px)44.8Align and Distill: Unifying and Improving Domain Adaptive Object Detection
ILLUME43.8To miss-attend is to misalign! Residual Self-Attentive Feature Alignment for Adapting Object Detectors-
ViSGA43.3Seeking Similarities over Differences: Similarity-based Domain Alignment for Adaptive Object Detection
SSA-DA42.5Bi-Dimensional Feature Alignment for Cross-Domain Object Detection-
DDF42.3Decompose to Adapt: Cross-domain Object Detection via Feature Disentanglement
DA-RetinaNet41.87An Unsupervised Domain Adaptation Scheme for Single-Stage Artwork Recognition in Cultural Sites
0 of 26 row(s) selected.
Unsupervised Domain Adaptation On Cityscapes 1 | SOTA | HyperAI超神经