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

Showing 101125 of 2454 papers

TitleStatusHype
CroCo: Self-Supervised Pre-training for 3D Vision Tasks by Cross-View CompletionCode2
Joint 2D-3D Multi-Task Learning on Cityscapes-3D: 3D Detection, Segmentation, and Depth EstimationCode2
GeoMVSNet: Learning Multi-View Stereo With Geometry PerceptionCode2
BEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object DetectionCode2
GeoBench: Benchmarking and Analyzing Monocular Geometry Estimation ModelsCode2
MultiMAE: Multi-modal Multi-task Masked AutoencodersCode2
MUTR3D: A Multi-camera Tracking Framework via 3D-to-2D QueriesCode2
GPS-Gaussian: Generalizable Pixel-wise 3D Gaussian Splatting for Real-time Human Novel View SynthesisCode2
BEVStereo: Enhancing Depth Estimation in Multi-view 3D Object Detection with Dynamic Temporal StereoCode2
Neural Ray Surfaces for Self-Supervised Learning of Depth and Ego-motionCode2
A Unified Image-Dense Annotation Generation Model for Underwater ScenesCode2
OPEN: Object-wise Position Embedding for Multi-view 3D Object DetectionCode2
HoloTime: Taming Video Diffusion Models for Panoramic 4D Scene GenerationCode2
PanoDreamer: Optimization-Based Single Image to 360 3D Scene With DiffusionCode2
Kick Back & Relax++: Scaling Beyond Ground-Truth Depth with SlowTV & CribsTVCode2
Atlas: End-to-End 3D Scene Reconstruction from Posed ImagesCode2
Boost 3D Reconstruction using Diffusion-based Monocular Camera CalibrationCode2
Physical 3D Adversarial Attacks against Monocular Depth Estimation in Autonomous DrivingCode2
360MonoDepth: High-Resolution 360deg Monocular Depth EstimationCode2
Few-shot Novel View Synthesis using Depth Aware 3D Gaussian SplattingCode2
Enforcing geometric constraints of virtual normal for depth predictionCode2
Consistent Video Depth EstimationCode2
Event-based Stereo Depth Estimation: A SurveyCode2
Deep3D: Fully Automatic 2D-to-3D Video Conversion with Deep Convolutional Neural NetworksCode2
EndoDAC: Efficient Adapting Foundation Model for Self-Supervised Depth Estimation from Any Endoscopic CameraCode2
Show:102550
← PrevPage 5 of 99Next →

Benchmark Results

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