List your company
ResearchFormerly WizWand

Boundary Detection on NYUD v2

80.31ODS F-measureTIGER-L
68.391671.485874.5877.6742Feb 2, 2026Feb 24, 2026Mar 18, 2026Apr 9, 2026May 1, 2026May 23, 2026Jun 14, 2026
Updated 4mo ago

Evaluation results

MethodLinks
TIGER-LBackbone=ViT-L2026.0680.31
TIFBackbone=Swin-L, Params(M)=32.912026.0479.52
SAKCategory=Conventional Multi-task Methods2026.0679.43
MTMamba++Backbone=Swin-Large, Decoder Pattern=Mamba-based decoder2026.0279.4
TIGER-BBackbone=ViT-B2026.0679
MTMambaBackbone=Swin-Large, Decoder Pattern=Mamba-based decoder2026.0278.7
MTMambaCategory=Conventional Multi-task Methods2026.0678.7
UV-M3TLBackbone=ViT-large2026.0278.68
TaskDiffusionBackbone=ViT-large, Decoder Pattern=Diffusion-based decoder2026.0278.64
MTLSI-NetBackbone=Swin-L, Params(M)=38.272026.0478.6
MLoREBackbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0278.43
MLoRECategory=Conventional Multi-task Methods2026.0678.43
TaskExpertBackbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0278.4
TaskExpertBackbone=ViT-L2026.0478.4
TaskExpertCategory=Conventional Multi-task Methods2026.0678.4
SEMCategory=Conventional Multi-task Methods2026.0678.4
TaskPrompterCategory=Conventional Multi-task Methods2026.0678.3
TaskPrompterBackbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0278.2
MQTransformerBackbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0278.2
TaskPrompterBackbone=ViT-L, Params(M)=88.002026.0478.2
MQTransformerCategory=Conventional Multi-task Methods2026.0678.2
InvPTCategory=Conventional Multi-task Methods2026.0678.2
InvPTBackbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0278.1
InvPT++Backbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0278.1
InvPTBackbone=ViT-L, Params(M)=97.962026.0478.1
RADIOCategory=Conventional Multi-task Methods2026.0677.97
ATRCBackbone=HRNet48, Decoder Pattern=CNN-based decoder2026.0277.94
ATRCCategory=Conventional Multi-task Methods2026.0677.94
MTI-NetBackbone=HRNet48, Decoder Pattern=CNN-based decoder2026.0277.86
MTI-NetCategory=Conventional Multi-task Methods2026.0677.86
TSP-TransformerBackbone=ViT-large, Decoder Pattern=Transformer-based decoder2026.0277.5
DeMTBackbone=Swin-T, Params(M)=32.072026.0476.9
MTLSI-NetBackbone=Swin-T, Params(M)=17.452026.0476.6
PAD-NetBackbone=HRNet-18, Params(M)=5.022026.0476.5
PAPBackbone=ResNet, Decoder Pattern=CNN-based decoder2026.0276.42
PSDBackbone=ResNet, Decoder Pattern=CNN-based decoder2026.0276.42
PSDBackbone=HRNet-18, Params(M)=4.712026.0476.42
MTLSI-NetBackbone=HRNet-18, Params(M)=13.942026.0476.4
Cross-StitchDecoder Pattern=CNN-based decoder2026.0276.38
PAD-NetBackbone=HRNet18, Decoder Pattern=CNN-based decoder2026.0276.38
PAD-NetCategory=Conventional Multi-task Methods2026.0676.38
ATRCBackbone=HRNet-18, Params(M)=5.062026.0476.34
MQTransformerBackbone=Swin-T, Params(M)=32.912026.0476.2
InvPT++Backbone=Swin-T2026.0476.2
DINOv2-ViT-LBackbone=ViT-L2026.0675.94
SAM-ViT-BBackbone=ViT-B2026.0675.92
OWLv2-ViT-LBackbone=ViT-L2026.0673.45
CLIP-ViT-LBackbone=ViT-L2026.0668.85