List your company

Affine-invariant Depth Estimation on ScanNet

0.021AbsRelOurs CSTM_label (ViT-Large)
-0.462162.799176.06059.32183Mar 22, 2024Aug 7, 2024Dec 23, 2024May 10, 2025Sep 25, 2025Feb 10, 2026Jun 29, 2026

Evaluation results

MethodLinks
Ours CSTM_label (ViT-Large)Backbone=ViT-Large, #Params=306M, Pretrain=142M, #Data Train=16M, Zero-shot transfer=true2024.030.02199.3
Ours CSTM_label (ViT-giant)Backbone=ViT-giant, #Params=1011M, Pretrain=142M, #Data Train=16M, Zero-shot transfer=true2024.030.02299.4
Ours CSTM_label (ViT-Small)Backbone=ViT-Small, #Params=22M, Pretrain=142M, #Data Train=16M, Zero-shot transfer=true2024.030.03398.5
DepthAnythingBackbone=ViT-Large, #Params=142M, Pretrain=142M, #Data Train=63.5M, Zero-shot transfer=true2024.030.04298
MarigoldBackbone=Latent diffusion V2, #Params=899M, Pretrain=5B, #Data Train=74K, Zero-shot transfer=true2024.030.06495.1
Ours CSTM_image (ConvNeXt-Large)Backbone=ConvNeXt-Large, #Params=198M, Pretrain=14.2M, #Data Train=8M, Zero-shot transfer=true2024.030.07494.2
Ours CSTM_label (ConvNeXt-Large)Backbone=ConvNeXt-Large, #Params=198M, Pretrain=14.2M, #Data Train=8M, Zero-shot transfer=true2024.030.07494.1
OmnidataBackbone=ViT-Base, Pretrain=1.3M, #Data Train=12.2M, Zero-shot transfer=true2024.030.07793.5
DPT-largeBackbone=ViT-Large, Pretrain=1.3M, #Data Train=188K, Zero-shot transfer=true2024.030.0893.8
LeresBackbone=ResNeXt101, Pretrain=1.3M, #Data Train=354K, Zero-shot transfer=true2024.030.09591.2
DiverseDepthBackbone=ResNeXt50, #Params=25M, Pretrain=1.3M, #Data Train=320K, Zero-shot transfer=true2024.030.10888.2
MiDaSBackbone=ResNeXt101, #Params=88M, Pretrain=1.3M, #Data Train=2M, Zero-shot transfer=true2024.030.11188.6
HDNBackbone=ViT-Large, #Params=306M, Pretrain=1.3M, #Data Train=300K, Zero-shot transfer=true2024.030.22869.4
DepthAnything V2Training Data=62.6M, Zero-shot=true2026.034.297.8
DepthAnything V2Training Data=62.6M2026.034.297.8
Lotus-2Generative Model=×, MTL=×, Inference Steps=12026.064.297.6
DepthAnythingTraining Data=62.6M, Zero-shot=true2026.034.398.1
DepthAnythingTraining Data=62.6M2026.034.398.1
DepthAnythingTraining Data=62.6M, Pre-trained Stable Diffusion=false2025.094.398.1
DINOv3mode=zero-shot2025.094.498.1
FE2Emode=zero-shot2025.094.497.5
FE2ETraining Data=71K, Pre-trained Stable Diffusion=false2025.094.497.5
FE2EGenerative Model=✓, MTL=✓, Inference Steps=12026.064.497.5
Qwen-Imagemode=zero-shot2025.094.797.4
IrisTraining Data=59K + 100K, Zero-shot=true2026.03597.1
Iris (ours)*Training Data=59K + 100K, diffusion-based=true2026.03597.1
Lotus-D*Training Data=59K, Zero-shot=true2026.035.596.5
Lotus-D*Training Data=59K, diffusion-based=true2026.035.596.5
Lotus-DTraining Data=59K, Pre-trained Stable Diffusion=true2025.095.596.5
Diffusion-E2E-FT*Training Data=74K, Zero-shot=true2026.035.896.5
Diffusion-E2E-FT*Training Data=74K, diffusion-based=true2026.035.896.5
Diffusion-E2E-FTTraining Data=74K, Pre-trained Stable Diffusion=true2025.095.896.5
MUSE-dGenerative Model=×, MTL=✓, Inference Steps=12026.065.896.2
Lotus-G*Training Data=59K, Zero-shot=true2026.035.995.7
Lotus-G*Training Data=59K, diffusion-based=true2026.035.995.7
Lotus-GTraining Data=59K, Pre-trained Stable Diffusion=true2025.095.995.7
Lotus-gGenerative Model=✓, MTL=×, Inference Steps=12026.06696
Lotus-dGenerative Model=×, MTL=×, Inference Steps=12026.06696.3
GeoWizardGenerative Model=✓, MTL=✓, Inference Steps=12026.066.195.3
GenPerceptTraining Data=74K, Pre-trained Stable Diffusion=false2025.096.296.1
OrchidGenerative Model=✓, MTL=✓, Inference Steps=12026.066.395.8
GeoWizard*Training Data=280K, Zero-shot=true2026.036.495
Marigold*Training Data=74K, Zero-shot=true2026.036.495.2
GenPercept*Training Data=90K, Zero-shot=true2026.036.496.1
GeoWizard*Training Data=280K, diffusion-based=true2026.036.495
Marigold*Training Data=74K, diffusion-based=true2026.036.495.2
GenPercept*Training Data=90K, diffusion-based=true2026.036.496.1
GeoWizardTraining Data=280K, Pre-trained Stable Diffusion=false2025.096.495
MarigoldTraining Data=74K, Pre-trained Stable Diffusion=true2025.096.495.2
Marigold v1.1Training Data=74K, Pre-trained Stable Diffusion=true2025.096.695.3
Marigoldv1.1Generative Model=✓, MTL=×, Inference Steps=12026.066.695.3
Marigold(LCM)*Training Data=74K, Zero-shot=true2026.036.994.6
Marigold(LCM)*Training Data=74K, diffusion-based=true2026.036.994.6
MarigoldGenerative Model=✓, MTL=×, Inference Steps=12026.066.994.5
MUSE-gGenerative Model=✓, MTL=✓, Inference Steps=12026.066.994.3
OmnidataTraining Data=12.2M, Zero-shot=true2026.037.593.6
OmnidataTraining Data=12.2M2026.037.593.6
DiceptionGenerative Model=✓, MTL=✓, Inference Steps=12026.067.593.8
HDNTraining Data=300K, Zero-shot=true2026.03893.9
HDNTraining Data=300K2026.03893.9
DPTTraining Data=1.4M, Zero-shot=true2026.038.293.4
DPTTraining Data=1.4M2026.038.293.4
LeResTraining Data=354K, Zero-shot=true2026.039.191.7
LeResTraining Data=354K2026.039.191.7
DiverseDepthTraining Data=320K, Zero-shot=true2026.0310.988.2
DiverseDepthTraining Data=320K2026.0310.988.2
MiDaSTraining Data=2M, Zero-shot=true2026.0312.184.6
MiDaSTraining Data=2M2026.0312.184.6
MiDaSTraining Data=2M, Pre-trained Stable Diffusion=false2025.0912.184.6