AI Paper Atlas

Research map / Benchmarks / NYU Depth V2

NYU Depth V2 leaderboard

52 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on NYU Depth V2. Best RMSE: 0.075 (DualFocus, NeurIPS 2025).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

NYU Depth V2 · RMSE

18 papers, each with its best reported number for its own method; lower is better.

8 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

NYU Depth V2 · Abs Rel

16 papers, each with its best reported number for its own method; lower is better.

6 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

NYU Depth V2 · REL

14 papers, each with its best reported number for its own method; lower is better.

#REL ↓MethodPaperTopic
10.056InstanceDepthInstance-Level Video Depth in Groups Beyond Occlusions
ICCV 2025 · Table 3
Depth and stereo
20.074DDVM (samples=2)
Efficient U-Net
The Surprising Effectiveness of Diffusion Models for Optical Flow and Monocular Depth Estimation
NeurIPS 2023 · Table 3
Fast diffusion sampling
30.089NeW-CRFs + ordinal entropy
NeW-CRFs
Improving Deep Regression with Ordinal Entropy
ICLR 2023 · Depth estimation table
Recognition and long-tail
40.099LocalBins
EfficientNet-B5
LocalBins: Improving Depth Estimation by Learning Local Distributions
ECCV 2022 · Table 1
Depth and stereo
50.103AdaBins
EfficientNet B5 encoder
AdaBins: Depth Estimation Using Adaptive Bins
CVPR 2021 · NYU-Depth-v2 results table
CNN design
60.106TransDepth (AGD+ViT, full)
ResNet-50
Transformer-Based Attention Networks for Continuous Pixel-Wise Prediction
ICCV 2021 · Table 2
Attention and transformers
70.108Ours (DAV)Guiding Monocular Depth Estimation Using Depth-Attention Volume
ECCV 2020 · Table 1
Depth and stereo
80.11MINE N=64 (RGB supervision)MINE: Towards Continuous Depth MPI With NeRF for Novel View Synthesis
ICCV 2021 · Table 3
NeRF and radiance fields
90.121Ours (ResNet50-95K)
ResNet50
Multi-Scale Continuous CRFs as Sequential Deep Networks for Monocular Depth Estimation
CVPR 2017 · Table: NYU Depth V2 comparison
Depth and stereo
100.131ST-CLSTM + GAN
ResNet-18
Exploiting Temporal Consistency for Real-Time Video Depth Estimation
ICCV 2019 · Table 2/4/5
Video segmentation and tracking
4 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

NYU Depth V2 · RMS

11 papers, each with its best reported number for its own method; lower is better.

#RMS ↓MethodPaperTopic
10.202InstanceDepthInstance-Level Video Depth in Groups Beyond Occlusions
ICCV 2025 · Table 3
Depth and stereo
20.321NeW-CRFs + ordinal entropy
NeW-CRFs
Improving Deep Regression with Ordinal Entropy
ICLR 2023 · Depth estimation table
Recognition and long-tail
30.357LocalBins
EfficientNet-B5
LocalBins: Improving Depth Estimation by Learning Local Distributions
ECCV 2022 · Table 1
Depth and stereo
40.364AdaBins
EfficientNet B5 encoder
AdaBins: Depth Estimation Using Adaptive Bins
CVPR 2021 · NYU-Depth-v2 results table
CNN design
50.365TransDepth (AGD+ViT, full)
ResNet-50
Transformer-Based Attention Networks for Continuous Pixel-Wise Prediction
ICCV 2021 · Table 2
Attention and transformers
60.408SAPA-G
BTS R50
SAPA: Similarity-Aware Point Affiliation for Feature Upsampling
NeurIPS 2022 · Table: Depth estimation result
CNN features
70.412Ours (DAV)Guiding Monocular Depth Estimation Using Depth-Attention Volume
ECCV 2020 · Table 1
Depth and stereo
80.481PH-RegDeep Regression Representation Learning with Topology
ICML 2024 · Depth estimation table
Recognition and long-tail
90.571ST-CLSTM + GAN
ResNet-18
Exploiting Temporal Consistency for Real-Time Video Depth Estimation
ICCV 2019 · Table 2/4/5
Video segmentation and tracking
100.586Ours (ResNet50-95K)
ResNet50
Multi-Scale Continuous CRFs as Sequential Deep Networks for Monocular Depth Estimation
CVPR 2017 · Table: NYU Depth V2 comparison
Depth and stereo
1 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

NYU Depth V2 · mIoU

10 papers, each with its best reported number for its own method; higher is better.

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.