SOTAVerified

Depth Estimation

Depth Estimation is the task of measuring the distance of each pixel relative to the camera. Depth is extracted from either monocular (single) or stereo (multiple views of a scene) images. Traditional methods use multi-view geometry to find the relationship between the images. Newer methods can directly estimate depth by minimizing the regression loss, or by learning to generate a novel view from a sequence. The most popular benchmarks are KITTI and NYUv2. Models are typically evaluated according to a RMS metric.

Source: DIODE: A Dense Indoor and Outdoor DEpth Dataset

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 1–50 of 2454 papers

TitleStatusHype
VGGT: Visual Geometry Grounded TransformerCode11
Sapiens: Foundation for Human Vision ModelsCode9
When Do We Not Need Larger Vision Models?Code9
Depth Pro: Sharp Monocular Metric Depth in Less Than a SecondCode9
Depth Anything V2Code9
Depth Anything: Unleashing the Power of Large-Scale Unlabeled DataCode9
Perception Encoder: The best visual embeddings are not at the output of the networkCode8
pySLAM: An Open-Source, Modular, and Extensible Framework for SLAMCode7
Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal EstimationCode7
Lumina-mGPT: Illuminate Flexible Photorealistic Text-to-Image Generation with Multimodal Generative PretrainingCode7
Endo-4DGS: Endoscopic Monocular Scene Reconstruction with 4D Gaussian SplattingCode7
Marigold: Affordable Adaptation of Diffusion-Based Image Generators for Image AnalysisCode7
FoundationStereo: Zero-Shot Stereo MatchingCode7
DINOv2: Learning Robust Visual Features without SupervisionCode6
DUSt3R: Geometric 3D Vision Made EasyCode5
PatchRefiner: Leveraging Synthetic Data for Real-Domain High-Resolution Monocular Metric Depth EstimationCode5
Prompting Depth Anything for 4K Resolution Accurate Metric Depth EstimationCode5
DepthCrafter: Generating Consistent Long Depth Sequences for Open-world VideosCode5
ZoeDepth: Zero-shot Transfer by Combining Relative and Metric DepthCode5
MegaSaM: Accurate, Fast, and Robust Structure and Motion from Casual Dynamic VideosCode5
DepthSplat: Connecting Gaussian Splatting and DepthCode5
UniDepth: Universal Monocular Metric Depth EstimationCode5
GAPartManip: A Large-scale Part-centric Dataset for Material-Agnostic Articulated Object ManipulationCode5
UniDepthV2: Universal Monocular Metric Depth Estimation Made SimplerCode5
Video Depth Anything: Consistent Depth Estimation for Super-Long VideosCode5
FeatUp: A Model-Agnostic Framework for Features at Any ResolutionCode5
Distill Any Depth: Distillation Creates a Stronger Monocular Depth EstimatorCode4
DN-Splatter: Depth and Normal Priors for Gaussian Splatting and MeshingCode4
DepthFM: Fast Monocular Depth Estimation with Flow MatchingCode4
One Diffusion to Generate Them AllCode4
Metric3D: Towards Zero-shot Metric 3D Prediction from A Single ImageCode4
Depth Any Video with Scalable Synthetic DataCode4
MonSter: Marry Monodepth to Stereo Unleashes PowerCode4
When Does Perceptual Alignment Benefit Vision Representations?Code4
LBM: Latent Bridge Matching for Fast Image-to-Image TranslationCode4
LET-3D-AP: Longitudinal Error Tolerant 3D Average Precision for Camera-Only 3D DetectionCode4
Repurposing Diffusion-Based Image Generators for Monocular Depth EstimationCode4
Fine-Tuning Image-Conditional Diffusion Models is Easier than You ThinkCode4
Self-Supervised Geometry-Guided Initialization for Robust Monocular Visual OdometryCode4
Attention on the SphereCode4
PF3plat: Pose-Free Feed-Forward 3D Gaussian SplattingCode3
DROID-Splat: Combining end-to-end SLAM with 3D Gaussian SplattingCode3
ECoDepth: Effective Conditioning of Diffusion Models for Monocular Depth EstimationCode3
What Matters When Repurposing Diffusion Models for General Dense Perception Tasks?Code3
PatchFusion: An End-to-End Tile-Based Framework for High-Resolution Monocular Metric Depth EstimationCode3
Relative Pose Estimation through Affine Corrections of Monocular Depth PriorsCode3
Denoising Vision TransformersCode3
Depth Any Camera: Zero-Shot Metric Depth Estimation from Any CameraCode3
DreamScene4D: Dynamic Multi-Object Scene Generation from Monocular VideosCode3
DEFOM-Stereo: Depth Foundation Model Based Stereo MatchingCode3
Show:102550
← PrevPage 1 of 50Next →
All datasetsStanford2D3D PanoramicNYU-Depth V2DCMeBDthequeScanNetV2Cityscapes testDIODEKITTI 2015Mars DTM EstimationScanNet4D Light Field DatasetKITTI Eigen split

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1OmniDepthRMSE0.62—Unverified
2SphereDepthRMSE0.45—Unverified
3Jin et al.RMSE0.42—Unverified
4BiFuse with fusionRMSE0.41—Unverified
5HoHoNet (ResNet-101)RMSE0.38—Unverified
6PanoDepthRMSE0.37—Unverified
7BiFuse++RMSE0.37—Unverified
8UniFuse with fusionRMSE0.37—Unverified
9DisConvRMSE0.37—Unverified
10SliceNetRMSE0.37—Unverified
#ModelMetricClaimedVerifiedStatus
1A2JmAP8.61—Unverified
2PAD-NetRMS0.79—Unverified
3MS-CRFRMS0.59—Unverified
4DORNRMS0.51—Unverified
5FreeformRMS0.43—Unverified
6Optimized, freeformRMS0.43—Unverified
7VNLRMS0.42—Unverified
8BTSRMS0.41—Unverified
9TransDepth (AGD+ ViT)RMS0.37—Unverified
10AdaBinsRMS0.36—Unverified
#ModelMetricClaimedVerifiedStatus
1T2NetAbs Rel0.35—Unverified
2MIDASAbs Rel0.31—Unverified
3Bhattacharjee et al.Abs Rel0.25—Unverified
#ModelMetricClaimedVerifiedStatus
1T2NetAbs Rel0.49—Unverified
2MIDASAbs Rel0.42—Unverified
3Bhattacharjee et al.Abs Rel0.38—Unverified
#ModelMetricClaimedVerifiedStatus
1LeReSabsolute relative error0.1—Unverified
2DELTASabsolute relative error0.09—Unverified
3Distill Any Depthabsolute relative error0.04—Unverified
#ModelMetricClaimedVerifiedStatus
1SDC-DepthRMSE6.92—Unverified
2SwinMTLRMSE6.35—Unverified
#ModelMetricClaimedVerifiedStatus
1AIP-BrownDelta < 1.250.36—Unverified
2LeResDelta < 1.250.23—Unverified
#ModelMetricClaimedVerifiedStatus
1H-Net (Ours)Absolute relative error (AbsRel)0.09—Unverified
2H-Net (Ours) Full EigenAbsolute relative error (AbsRel)0.08—Unverified
#ModelMetricClaimedVerifiedStatus
1GLPDepthDelta < 1.250.43—Unverified
2SRDINET (Model A)Delta < 1.250.4—Unverified
#ModelMetricClaimedVerifiedStatus
1Atlas (finetuned)RMSE0.17—Unverified
2Atlas (plain)RMSE0.17—Unverified
#ModelMetricClaimedVerifiedStatus
1LFattNetBadPix(0.01)17.23—Unverified
#ModelMetricClaimedVerifiedStatus
1LightDepthNumber of parameters (M)42.6—Unverified
#ModelMetricClaimedVerifiedStatus
1UniFuseAbs Rel0.11—Unverified
#ModelMetricClaimedVerifiedStatus
1X-TC (Cross-Task Consistency)L1 error1.63—Unverified