Depth Completion Evaluation



The depth completion and depth prediction evaluation are related to our work published in Sparsity Invariant CNNs (THREEDV 2017). It
contains over 93 thousand depth maps with corresponding raw LiDaR scans and RGB images, aligned with the "raw data" of the KITTI dataset.
Given the large amount of training data, this dataset shall allow a training of complex deep learning models for the tasks of depth completion
and single image depth prediction. Also, we provide manually selected images with unpublished depth maps to serve as a benchmark for those
two challenging tasks.

Make sure to unzip annotated depth maps and raw LiDaR scans into the same directory so that all corresponding files end up in the same folder
structure. The structure of all provided depth maps is aligned with the structure of our raw data to easily find corresponding left and right images,
or other provided information.


Note: On 12.04.2018 we have fixed a small error in the file data_depth_velodyne.zip, please download this file again if you have an old version.


All methods providing less than 100 % density have been interpolated using simple background interpolation as explained in the corresponding header file in the development kit.

    Our evaluation table ranks all methods according to the root mean squared error (RMSE) of the inverse depth maps.
    However, we also provide some other metrics:
  • iRMSE:  Root mean squared error of the inverse depth [1/km]
  • iMAE:    Mean absolute error of the inverse depth [1/km]
  • RMSE:   Root mean squared error [mm]
  • MAE:     Mean absolute error [mm]


Important Policy Update: As more and more non-published work and re-implementations of existing work is submitted to KITTI, we have established a new policy: from now on, only submissions with significant novelty that are leading to a peer-reviewed paper in a conference or journal are allowed. Minor modifications of existing algorithms or student research projects are not allowed. Such work must be evaluated on a split of the training set. To ensure that our policy is adopted, new users must detail their status, describe their work and specify the targeted venue during registration. Furthermore, we will regularly delete all entries that are 6 months old but are still anonymous or do not have a paper associated with them. For conferences, 6 month is enough to determine if a paper has been accepted and to add the bibliography information. For longer review cycles, you need to resubmit your results.
Additional information used by the methods
  • Additional training data: Use of additional data sources for training (see details)
  • RGB image: Use of RGB images for depth completion

Method Setting Code iRMSE iMAE RMSE MAE Runtime Environment
1 DySPN 1.88 0.82 709.12 192.71 0.16 s 1 core @ 2.5 Ghz (C/C++)
2 Anonymous 2.03 0.90 709.41 205.49 0.01 s 1 core @ 2.5 Ghz (C/C++)
3 MCF 2.03 0.90 710.11 205.12 0.09 s GPU @ 2.5 Ghz (Python)
4 RigNet 2.08 0.90 712.66 203.25 0.20 s GPU @ 2.5 Ghz (Python)
Z. Yan, K. Wang, X. Li, Z. Zhang, B. Xu, J. Li and J. Yang: RigNet: Repetitive Image Guided Network for Depth Completion. 2021.
5 SEHLNet 2.07 0.91 714.71 207.51 0.07 s GPU @ 2.5 Ghz (Python)
6 RigNet_v1 2.19 0.98 717.88 216.74 0.20 s GPU @ 2.0 Ghz (Python)
7 MF 2.21 0.94 719.85 208.11 0.1 s GPU @ 2.5 Ghz (Python)
8 FC 2.35 0.91 722.01 205.44 0.1 s GPU @ 2.5 Ghz (Python)
9 DC 2.09 0.92 723.35 210.06 0.01 s 1 core @ 2.5 Ghz (C/C++)
10 SGSNet 2.11 0.92 723.67 209.54 0.02 s GPU @ 2.5 Ghz (Python)
11 NNNet 1.99 0.88 724.14 205.57 0.03 s GPU @ 2.0 Ghz (Python)
12 Test 2.08 0.90 726.71 207.25 0.01 s 1 core @ 2.5 Ghz (C/C++)
13 3DCGNet 2.26 1.00 727.40 219.10 0.14 s GPU @ 1.5 Ghz (Python)
14 NLSANet 2.28 0.98 728.31 216.46 0.15 s GPU @ 1.5 Ghz (Python)
15 genet 2.33 1.03 728.44 219.50 0.15 s 1 core @ 2.5 Ghz (Python)
16 DC_BB_SMT 2.15 0.95 728.47 213.33 1 s 1 core @ 2.5 Ghz (C/C++)
17 DepthNet 2.05 0.91 728.93 208.86 0.03 s GPU @ 2.5 Ghz (Python)
18 SCT 2.15 0.93 729.11 210.01 0.02 s GPU @ 2.0 Ghz (Python)
19 DenseLConv-64 2.10 0.93 729.88 210.06 0.12 s GPU @ 2.5 Ghz (Python)
20 PENet code 2.17 0.94 730.08 210.55 0.032s GPU @ 2.5 Ghz (Python)
M. Hu, S. Wang, B. Li, S. Ning, L. Fan and X. Gong: PENet: Towards Precise and Efficient Image Guided Depth Completion. ICRA 2021.
21 SN 2.04 0.85 730.72 200.65 0.03 s 1 core @ 2.5 Ghz (Python)
22 dd 2.20 1.03 731.53 219.23 dd s 1 core @ 2.5 Ghz (C/C++)
23 DC_BB 2.07 0.92 731.97 211.84 0.1 s 1 core @ 2.5 Ghz (C/C++)
24 CFN 2.12 0.93 732.08 211.21 0.05 s 1 core @ 2.5 Ghz (Python)
25 MFCFR-Net 2.19 0.98 732.72 216.75 0.13 s GPU @ 2.5 Ghz (Python)
26 ACMNet code 2.08 0.90 732.99 206.80 0.08 s 1 core @ 2.5 Ghz (Python + C/C++)
S. Zhao, M. Gong, H. Fu and D. Tao: Adaptive context-aware multi-modal network for depth completion. IEEE Transactions on Image Processing 2021.
27 SPL 2.09 0.93 733.44 212.49 0.03 s 1 core @ 2.5 Ghz (Python)
28 FMPN 2.11 0.92 733.69 211.15 0.015 s 1 core @ 2.5 Ghz (Python)
29 FCFR-Net 2.20 0.98 735.81 217.15 0.13 s GPU @ 2.5 Ghz (Python)
L. Liu, X. Song, X. Lyu, J. Diao, M. Wang, Y. Liu and L. Zhang: FCFR-Net: Feature Fusion based Coarse- to-Fine Residual Learning for Depth Completion. Proceedings of the AAAI Conference on Artificial Intelligence 2021.
30 DC 2.15 0.95 736.09 214.77 0.01 s 1 core @ 2.5 Ghz (C/C++)
31 GuideNet code 2.25 0.99 736.24 218.83 0.14 s GPU @ 1.5 Ghz (Python + C/C++)
J. Tang, F. Tian, W. Feng, J. Li and P. Tan: Learning Guided Convolutional Network for Depth Completion. IEEE Transactions on Image Processing(TIP) 2020.
32 ADLAB-MDANet 2.12 0.99 738.23 214.99 0.04 s GPU @ 2.5 Ghz (Python)
33 AdaptiveFusion 2.28 0.97 740.16 215.69 0.05 s GPU @ 1.5 Ghz (Python)
34 CMPN 1.93 0.86 740.91 203.48 0.2 s 1 core @ 2.5 Ghz (C/C++)
35 ENet code 2.14 0.95 741.30 216.26 0.019 s GPU @ 2.5 Ghz (Python)
M. Hu, S. Wang, B. Li, S. Ning, L. Fan and X. Gong: PENet: Towards Precise and Efficient Image Guided Depth Completion. ICRA 2021.
36 NLSPN code 1.99 0.84 741.68 199.59 0.22 s GPU @ 1.5 Ghz (Python)
J. Park, K. Joo, Z. Hu, C. Liu and I. Kweon: Non-Local Spatial Propagation Network for Depth Completion. European Conference on Computer Vision (ECCV) 2020.
37 Point-Fusion 1.97 0.85 741.94 201.07 0.081 s 1 core @ 2.5 Ghz (Python)
38 try1 2.15 0.94 743.46 213.41 0.1 s 1 core @ 2.5 Ghz (C/C++)
39 CSPN++ 2.07 0.90 743.69 209.28 0.2 s 1 core @ 2.5 Ghz (C/C++)
X. Cheng, P. Wang, G. Chenye and R. Yang: CSPN++: Learning Context and Resource Aware Convolutional Spatial Propagation Networks for Depth Completion. Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI-20) 2020.
40 ACMNet code 2.08 0.90 744.91 206.09 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
S. Zhao, M. Gong, H. Fu and D. Tao: Adaptive context-aware multi-modal network for depth completion. IEEE Transactions on Image Processing 2021.
41 RG-DSPN 2.02 0.85 745.48 197.63 0.2 s GPU @ 1.5 Ghz (Python)
42 MCDnet 2.21 0.96 747.09 215.47 0.04 s GPU @ 1.5 Ghz (Python)
43 SPKD-MSG-CHN_64 2.16 0.92 750.90 212.66 0.04 s 1 core @ 2.5 Ghz (Python)
44 Ms_Unc_UARes-B code 1.98 0.85 751.59 198.09 0.1 s GPU @ 2.5 Ghz (Python)
Y. Zhu, W. Dong, L. Li, J. Wu, X. Li and G. Shi: Robust Depth Completion with Uncertainty-Driven Loss Functions. accepted by AAAI2022 .
45 UberATG-FuseNet 2.34 1.14 752.88 221.19 0.09 s GPU @ 2.5 Ghz (Python)
Y. Chen, B. Yang, M. Liang and R. Urtasun: Learning Joint 2D-3D Representations for Depth Completion. ICCV 2019.
46 DenseLConv-32 2.22 0.94 753.78 216.15 0.06 s GPU @ 2.5 Ghz (Python)
47 DenseLiDAR 2.25 0.96 755.41 214.13 0.02 s 1 core @ 2.5 Ghz (Python)
J. Gu, Z. Xiang, Y. Ye and L. Wang: DenseLiDAR: A Real-Time Pseudo Dense Depth Guided Depth Completion Network. IEEE Robotics and Automation Letters 2021.
48 DeepLiDAR code 2.56 1.15 758.38 226.50 0.07s GPU @ 1.5 Ghz (Python)
J. Qiu, Z. Cui, Y. Zhang, X. Zhang, S. Liu, B. Zeng and M. Pollefeys: DeepLiDAR: Deep Surface Normal Guided Depth Prediction for Outdoor Scene From Sparse LiDAR Data and Single Color Image. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2019.
49 DANConv code 2.17 0.92 759.65 213.68 0.05 s GPU @ 2.5 Ghz (Python)
L. Yan, K. Liu and G. Long: DAN-Conv: Depth aware non-local convolution for LiDAR depth completion. Electronics Letters 2021.
50 MSG-CHN code 2.30 0.98 762.19 220.41 0.01 s GPU @ 2.5 Ghz (Python + C/C++)
A. Li, Z. Yuan, Y. Ling, W. Chi, C. Zhang and others: A Multi-Scale Guided Cascade Hourglass Network for Depth Completion. The IEEE Winter Conference on Applications of Computer Vision 2020.
51 ABCD code 2.29 0.97 764.61 220.86 0.02 s 1 core @ 2.5 Ghz (C/C++)
Y. Jeon, H. Kim and S. Seo: ABCD: Attentive Bilateral Convolutional Network for Robust Depth Completion. IEEE Robotics and Automation Letters 2021.
52 DSPN 2.47 1.03 766.74 220.36 0.34 s 1 core @ 2.5 Ghz (Python)
Z. Xu, H. Yin and J. Yao: Deformable Spatial Propagation Networks For Depth Completion. 2020 IEEE International Conference on Image Processing (ICIP) 2020.
53 MsG 2.41 1.17 768.35 235.55 0.24 s GPU @ 2.5 Ghz (Python)
54 GAE-Net 2.18 0.93 769.06 214.99 0.05 s GPU @ 2.5 Ghz (Python)
55 RGB_guide&certainty code 2.19 0.93 772.87 215.02 0.02 s GPU @ 1.5 Ghz (Python)
W. Van Gansbeke, D. Neven, B. De Brabandere and L. Van Gool: Sparse and noisy LiDAR completion with RGB guidance and uncertainty. International Conference on Machine Vision Applications (MVA) 2019.
56 MS 2.54 1.20 774.67 239.04 0.24 s GPU @ 2.5 Ghz (Python)
57 DVMN 2.21 0.94 776.31 220.37 0.12 s GPU @ 1.5 Ghz (Python)
L. Reichardt, P. Mangat and O. Wasenmüller: DVMN: Dense Validity Mask Network for Depth Completion. IEEE International Conference on Intelligent Transportation (ITSC) 2021.
58 PwP 2.42 1.13 777.05 235.17 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
H. Yan Xu: Depth Completion from Sparse LiDAR Data with Depth-Normal Constraints. Proceedings of the IEEE International Conference on Computer Vision 2019.
59 DenseTeacher 2.08 0.89 778.96 211.41 0.1 s 1 core @ 2.5 Ghz (Python)
60 ksnet 2.42 1.08 783.56 239.57 15 s 1 core @ 2.5 Ghz (C/C++)
61 Revisiting code 2.42 0.99 792.80 225.81 0.05 s GPU @ 2.0 Ghz (Python)
L. Yan, K. Liu and E. Belyaev: Revisiting Sparsity Invariant Convolution: A Network for Image Guided Depth Completion. IEEE Access 2020.
62 Ms_Unc_UARes code 1.98 0.83 795.61 190.88 0.08 s GPU @ 2.5 Ghz (Python)
Y. Zhu, W. Dong, L. Li, J. Wu, X. Li and G. Shi: Robust Depth Completion with Uncertainty-Driven Loss Functions. accepted by AAAI2022 .
63 CrossGuidance 2.73 1.33 807.42 253.98 0.2 s 1 core @ 2.5 Ghz (Python)
S. Lee, J. Lee, D. Kim and J. Kim: Deep Architecture with Cross Guidance Between Single Image and Sparse LiDAR Data for Depth Completion. IEEE Access 2020.
64 Sparse-to-Dense (gd) code 2.80 1.21 814.73 249.95 0.08 s GPU @ 1.5 Ghz (Python)
F. Ma, G. Cavalheiro and S. Karaman: Self-supervised Sparse-to-Dense: Self- supervised Depth Completion from LiDAR and Monocular Camera. 2019 IEEE International Conference on Robotics and Automation (ICRA) 2019.
65 MonDi 2.21 0.96 815.64 231.07 0.01 s 1 core @ 1.5 Ghz (C/C++)
66 DYPN 2.11 0.91 821.01 219.31 0.58 s GPU @ 1.5 Ghz (Python)
67 NConv-CNN-L2 (gd) code 2.60 1.03 829.98 233.26 0.02 s GPU @ 1.5 Ghz (Python)
A. Eldesokey, M. Felsberg and F. Khan: Confidence propagation through cnns for guided sparse depth regression. IEEE transactions on pattern analysis and machine intelligence 2019.
68 DDP 2.10 0.85 832.94 203.96 0.08 s GPU @ 1.5 Ghz (Python)
Y. Yang, A. Wong and S. Soatto: Dense depth posterior (ddp) from single image and sparse range. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2019.
69 SSGP 2.51 1.09 838.22 244.70 0.14 s RTX 2080 Ti
R. Schuster, O. Wasenmüller, C. Unger and D. Stricker: SSGP: Sparse Spatial Guided Propagation for Robust and Generic Interpolation. IEEE Winter Conference on Applications of Computer Vision (WACV) 2021.
70 HiReNet 2.38 0.99 842.48 229.69 0.18 s GPU @ 2.5 Ghz (Python)
ERROR: Wrong syntax in BIBTEX file.
71 ScaffFusion-SSL code 3.24 0.88 847.22 205.75 0.03 s 1 core @ 1.5 Ghz (Python)
A. Wong, S. Cicek and S. Soatto: Learning topology from synthetic data for unsupervised depth completion. IEEE Robotics and Automation Letters 2021.
72 NConv-CNN-L1 (gd) code 2.52 0.92 859.22 207.77 0.02 s GPU @ 1.5 Ghz (Python)
A. Eldesokey, M. Felsberg and F. Khan: Confidence propagation through cnns for guided sparse depth regression. IEEE transactions on pattern analysis and machine intelligence 2019.
73 BDBF 2.37 0.89 900.38 216.44 0.15 s 1 core @ 3.0 Ghz (Python)
74 IR_L2 4.92 1.35 901.43 292.36 0.05 s GPU @ 2.5 Ghz (Python)
K. Lu, N. Barnes, S. Anwar and L. Zheng: From Depth What Can You See? Depth Completion via Auxiliary Image Reconstruction. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2020.
75 Spade-RGBsD 2.17 0.95 917.64 234.81 0.07 s GPU @ 2.5 Ghz (Python)
M. Jaritz, R. Charette, E. Wirbel, X. Perrotton and F. Nashashibi: Sparse and Dense Data with CNNs: Depth Completion and Semantic Segmentation. International Conference on 3D Vision (3DV) 2018.
76 DUNet code 2.69 1.04 917.76 244.36 0.11 s 1 core @ 2.5 Ghz (Python)
77 glob_guide&certainty code 2.80 1.07 922.93 249.11 0.02 s GPU @ 1.5 Ghz (Python)
W. Van Gansbeke, D. Neven, B. De Brabandere and L. Van Gool: Sparse and noisy LiDAR completion with RGB guidance and uncertainty. International Conference on Machine Vision Applications (MVA) 2019.
78 DNet code 2.71 1.05 923.13 246.59 15 s 1 core @ 2.5 Ghz (C/C++)
79 SNet code 2.95 1.13 933.59 257.81 15 s 1 core @ 2.5 Ghz (C/C++)
80 LDC 2.80 1.04 936.26 248.56 0.04 s GPU @ 2.5 Ghz (Python)
81 DFineNet code 3.21 1.39 943.89 304.17 0.02 s GPU @ 2.5 Ghz (Python)
Y. Zhang, T. Nguyen, I. Miller, S. Shivakumar, S. Chen, C. Taylor and V. Kumar: DFineNet: Ego-Motion Estimation and Depth Refinement from Sparse, Noisy Depth Input with RGB Guidance. CoRR 2019.
82 Sparse-to-Dense (d) code 3.21 1.35 954.36 288.64 0.04 s GPU @ 1.5 Ghz (Python)
F. Ma, G. Cavalheiro and S. Karaman: Self-supervised Sparse-to-Dense: Self- supervised Depth Completion from LiDAR and Monocular Camera. 2019 IEEE International Conference on Robotics and Automation (ICRA) 2019.
83 SIUNet_L2 2.88 1.10 958.91 256.19 0.01 s GPU @ 2.5 Ghz (Python)
84 pNCNN (d) code 3.37 1.05 960.05 251.77 0.02 s 1 core @ 2.5 Ghz (Python)
A. Eldesokey, M. Felsberg, K. Holmquist and M. Persson: Uncertainty-Aware CNNs for Depth Completion: Uncertainty from Beginning to End. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2020.
85 Conf-Net code 3.10 1.09 962.28 257.54 0.02 s GPU @ 2.5 Ghz (Python)
H. Hekmatian, S. Al-Stouhi and J. Jin: Conf-Net: Predicting Depth Completion Error-Map For High-Confidence Dense 3D Point- Cloud. 2019.
86 DCrgb_80b_3coef 2.43 0.98 965.87 215.75 0.15 s 1 core @ 2.5 Ghz (C/C++)
S. Imran, Y. Long, X. Liu and D. Morris: Depth coefficients for depth completion. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2019.
87 UDC_delving 3.05 1.21 972.70 273.82 0.01 s GPU @ 2.5 Ghz (Python)
88 DCd_all 2.87 1.13 988.38 252.21 0.1 s 1 core @ 2.5 Ghz (C/C++)
S. Imran, Y. Long, X. Liu and D. Morris: Depth coefficients for depth completion. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2019.
89 LW-DepthNet 2.99 1.09 991.88 261.67 0.09 s GPU @ 2.5 Ghz (Python)
L. Bai, Y. Zhao, M. Elhousni and X. Huang: DepthNet: Real-Time LiDAR Point Cloud Depth Completion for Autonomous Vehicles. arXiv preprint arXiv:2007.02438 2020.
90 CSPN 2.93 1.15 1019.64 279.46 1 s GPU @ 2.5 Ghz (Python + C/C++)
X. Cheng, P. Wang and R. Yang: Depth estimation via affinity learned with convolutional spatial propagation network. Proceedings of the European Conference on Computer Vision (ECCV) 2018.
X. Cheng, P. Wang and R. Yang: Learning Depth with Convolutional Spatial Propagation Network. arXiv preprint arXiv:1810.02695 2018.
91 SIUNet_L1 2.73 0.96 1026.61 227.28 0.01 s GPU (Python)
92 Spade-sD 2.60 0.98 1035.29 248.32 0.04 s GPU @ 2.5 Ghz (Python)
M. Jaritz, R. Charette, E. Wirbel, X. Perrotton and F. Nashashibi: Sparse and Dense Data with CNNs: Depth Completion and Semantic Segmentation. International Conference on 3D Vision (3DV) 2018.
93 Morph-Net 3.84 1.57 1045.45 310.49 0.17 s GPU @ 1.5 Ghz (Matlab + C/C++)
M. Dimitrievski, P. Veelaert and W. Philips: Learning morphological operators for depth completion. Advanced Concepts for Intelligent Vision Systems 2018.
94 SynthProjV 3.12 1.13 1062.48 268.37 0.1 s 1 core @ 2.5 Ghz (C/C++)
A. Lopez-Rodriguez, B. Busam and K. Mikolajczyk: Project to Adapt: Domain Adaptation for Depth Completion from Noisy and Sparse Sensor Data. Asian Conference on Computer Vision (ACCV) 2020.
95 KBNet code 2.95 1.02 1069.47 256.76 0.01 s 1 core @ 2.5 Ghz (C/C++)
A. Wong and S. Soatto: Unsupervised Depth Completion with Calibrated Backprojection Layers. Proceedings of the IEEE International Conference on Computer Vision (ICCV) 2021.
96 3DRGB-1 3.03 1.03 1072.39 262.85 0.01 s 1 core @ 1.5 Ghz (Python)
ERROR: Wrong syntax in BIBTEX file.
97 VLW-DepthNet 3.43 1.21 1077.22 282.02 0.09 GPU @ 2.5 Ghz (Python)
L. Bai, Y. Zhao, M. Elhousni and X. Huang: DepthNet: Real-Time LiDAR Point Cloud Depth Completion for Autonomous Vehicles. arXiv preprint arXiv:2007.02438 2020.
98 3DRGB 3.10 1.07 1087.50 271.79 0.01 s 1 core @ 1.5 Ghz (Python)
99 SynthProj 3.53 1.19 1095.26 280.42 0.1 s 1 core @ 2.5 Ghz (C/C++)
A. Lopez-Rodriguez, B. Busam and K. Mikolajczyk: Project to Adapt: Domain Adaptation for Depth Completion from Noisy and Sparse Sensor Data. Asian Conference on Computer Vision (ACCV) 2020.
100 DCd_3 2.95 1.07 1109.04 234.01 0.1 s 1 core @ 2.5 Ghz (C/C++)
S. Imran, Y. Long, X. Liu and D. Morris: Depth coefficients for depth completion. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2019.
101 ScaffFusion code 3.32 1.17 1121.89 282.86 0.03 s 1 core @ 1.5 Ghz (Python)
A. Wong, S. Cicek and S. Soatto: Learning topology from synthetic data for unsupervised depth completion. IEEE Robotics and Automation Letters 2021.
102 AdaFrame-VGG8 code 3.32 1.16 1125.67 291.62 0.02 s GPU @ 1.5 Ghz (Python)
A. Wong, X. Fei, B. Hong and S. Soatto: An Adaptive Framework for Learning Unsupervised Depth Completion. IEEE Robotics and Automation Letters 2021.
103 VOICED code 3.56 1.20 1169.97 299.41 0.02 s 1 core @ 2.5 Ghz (C/C++)
A. Wong, X. Fei, S. Tsuei and S. Soatto: Unsupervised Depth Completion from Visual Inertial Odometry. IEEE Robotics and Automation Letters 2020.
104 DFuseNet code 3.62 1.79 1206.66 429.93 0.08 s GPU @ 2.0 Ghz (C/C++)
S. Shivakumar, T. Nguyen, S. Chen and C. Taylor: DFuseNet: Deep Fusion of RGB and Sparse Depth Information for Image Guided Dense Depth Completion. arXiv preprint arXiv:1902.00761 2019.
105 NonLearning Complete 3.63 1.23 1222.00 303.82 0.84 s 1 core @ 3.5 Ghz (Python)
B. Krauss, G. Schroeder, M. Gustke and A. Hussein: Deterministic Guided LiDAR Depth Map Completion. 2021 IEEE Intelligent Vehicles Symposium (IV) 2021.
106 PDC 3.89 1.26 1227.96 288.55 10 s 1 core @ 2.5 Ghz (Python)
D. Teutscher, P. Mangat and O. Wasenmüller: PDC: Piecewise Depth Completion utilizing Superpixels. IEEE International Conference on Intelligent Transportation (ITSC) 2021.
107 Physical_Surface_Mod code 3.76 1.21 1239.84 298.30 0.06 s 1 core @ 2.5 Ghz (C/C++)
Y. Zhao, L. Bai, Z. Zhang and X. Huang: A Surface Geometry Model for LiDAR Depth Completion. IEEE Robotics and Automation Letters 2021.
108 NG_Depth code 14.93 1.38 1266.22 305.98 0.8 s 1 core @ 2.5 Ghz (C/C++)
P. An, Y. Gao, W. Fu, J. Ma, B. Fang and K. Yu: Lambertian Model Based Normal Guided Depth Completion for LiDAR-Camera System. IEEE GRSL 2021.
109 NConv-CNN (d) code 4.67 1.52 1268.22 360.28 0.01 s GPU @ 1.5 Ghz (Python)
A. Eldesokey, M. Felsberg and F. Khan: Propagating Confidences through CNNs for Sparse Data Regression. 2018.
110 IP-Basic code 3.78 1.29 1288.46 302.60 0.011 s 1 core @ >3.5 Ghz (Python)
J. Ku, A. Harakeh and S. Waslander: In Defense of Classical Image Processing: Fast Depth Completion on the CPU. 2018 15th Conference on Computer and Robot Vision (CRV) 2018.
111 Sparse2Dense(w/o gt) code 4.07 1.57 1299.85 350.32 0.08 s GPU @ 1.5 Ghz (Python + C/C++)
F. Ma, G. Cavalheiro and S. Karaman: Self-supervised Sparse-to-Dense: Self- supervised Depth Completion from LiDAR and Monocular Camera. 2019 IEEE International Conference on Robotics and Automation (ICRA) 2019.
112 ADNN code 59.39 3.19 1325.37 439.48 .04 s GPU @ 2.5 Ghz (Python)
S. Nathaniel Chodosh: Deep Convolutional Compressed Sensing for LiDAR Depth Completion. Asian Conference on Computer Vision (ACCV) 2018.
113 NN+CNN 3.25 1.29 1419.75 416.14 0.02 s GPU
J. Uhrig, N. Schneider, L. Schneider, U. Franke, T. Brox and A. Geiger: Sparsity Invariant CNNs. International Conference on 3D Vision (3DV) 2017.
114 B-ADT 4.16 1.23 1480.36 298.72 0.120 sec. GPU
Y. Yao, M. Roxas, R. Ishikawa, S. Ando, j. shimamura and T. Oishi: Discontinuous and Smooth Depth Completion with Binary Anisotropic Diffusion Tensor. IEEE Robotics and Automation Letters 2020.
115 SparseConvs code 4.94 1.78 1601.33 481.27 0.01 s GPU
J. Uhrig, N. Schneider, L. Schneider, U. Franke, T. Brox and A. Geiger: Sparsity Invariant CNNs. International Conference on 3D Vision (3DV) 2017.
116 SIUNet_L1_Zero-shot 5.49 1.55 1613.10 361.46 0.01 s GPU @ 2.5 Ghz (Python)
117 NadarayaW 6.34 1.84 1852.60 416.77 0.05 s 1 core @ 2.5 Ghz (Python)
J. Uhrig, N. Schneider, L. Schneider, U. Franke, T. Brox and A. Geiger: Sparsity Invariant CNNs. International Conference on 3D Vision (3DV) 2017.
118 SGDU 7.38 2.05 2312.57 605.47 0.2 s 4 cores @ 2.5 Ghz (C/C++)
N. Schneider, L. Schneider, P. Pinggera, U. Franke, M. Pollefeys and C. Stiller: Semantically Guided Depth Upsampling. German Conference on Pattern Recognition 2016.
Table as LaTeX | Only published Methods




Related Datasets

  • SYNTHIA Dataset: SYNTHIA is a collection of photo-realistic frames rendered from a virtual city and comes with precise pixel-level semantic annotations as well as pixel-wise depth information. The dataset consists of +200,000 HD images from video streams and +20,000 HD images from independent snapshots.
  • Middlebury Stereo Evaluation: The classic stereo evaluation benchmark, featuring four test images in version 2 of the benchmark, with very accurate ground truth from a structured light system. 38 image pairs are provided in total.
  • Make3D Range Image Data: Images with small-resolution ground truth used to learn and evaluate depth from single monocular images.
  • Virtual KITTI Dataset: Virtual KITTI contains 50 high-resolution monocular videos (21,260 frames) generated from five different virtual worlds in urban settings under different imaging and weather conditions.
  • Scene Flow Dataset: The Freiburg Scene Flow Dataset collection has been used to train convolutional networks for disparity, optical flow, and scene flow estimation. The collection contains more than 39000 stereo frames in 960x540 pixel resolution, rendered from various synthetic sequences.

Citation

When using this dataset in your research, we will be happy if you cite us:
@INPROCEEDINGS{Uhrig2017THREEDV,
  author = {Jonas Uhrig and Nick Schneider and Lukas Schneider and Uwe Franke and Thomas Brox and Andreas Geiger},
  title = {Sparsity Invariant CNNs},
  booktitle = {International Conference on 3D Vision (3DV)},
  year = {2017}
}



eXTReMe Tracker