Semi Supervised Semantic Segmentation On 22

Metriken

Validation mIoU

Ergebnisse

Leistungsergebnisse verschiedener Modelle zu diesem Benchmark

		Paper Title
UniMatch V2 (DINOv2-B)	83.6	UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation
SemiVL (ViT-B/16)	77.9	SemiVL: Semi-Supervised Semantic Segmentation with Vision-Language Guidance
PrevMatch (ResNet-101)	77.7%	Revisiting and Maximizing Temporal Knowledge in Semi-supervised Semantic Segmentation
CorrMatch (Deeplabv3+ with ResNet-101)	77.3	CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentation
S4MC	77.00	Semi-Supervised Semantic Segmentation via Marginal Contextual Information
Dual Teacher (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)	76.81	Switching Temporary Teachers for Semi-Supervised Semantic Segmentation
UniMatch (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)	76.59	Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation
n-CPS (ResNet-50)	76.08	n-CPS: Generalising Cross Pseudo Supervision to n Networks for Semi-Supervised Semantic Segmentation
AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)	75.83%	Semi-Supervised Semantic Segmentation via Adaptive Equalization Learning
PrevMatch (ResNet-50)	75.8%	Revisiting and Maximizing Temporal Knowledge in Semi-supervised Semantic Segmentation
CW-BASS (DeepLab v3+ with ResNet-50)	75.00	CW-BASS: Confidence-Weighted Boundary-Aware Learning for Semi-Supervised Semantic Segmentation
U2PL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K, AEL)	74.90%	Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels
PCR (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)	73.41%	Semi-supervised Semantic Segmentation with Prototype-based Consistency Regularization
Dense FixMatch (DeepLabv3+ ResNet-101, over-sampling, single pass eval)	71.1%	Dense FixMatch: a simple semi-supervised learning method for pixel-wise prediction tasks
Dense FixMatch (DeepLabv3+ ResNet-50, uniform sampling, single pass eval)	70.65%	Dense FixMatch: a simple semi-supervised learning method for pixel-wise prediction tasks
CPCL (DeepLab v3+ with ResNet-50)	69.92%	Conservative-Progressive Collaborative Learning for Semi-supervised Semantic Segmentation
CPS (DeepLab v3+ with ResNet-101)	69.8	Semi-Supervised Semantic Segmentation with Cross Pseudo Supervision

0 of 17 row(s) selected.