This table lists the benchmark results for the high-res multi-view scenario.
The following metrics are evaluated:
Accuracy [%]: The fraction of the reconstruction which is closer to the ground truth than the evaluation threshold distance (*). Larger is better.Completeness [%]: The fraction of the ground truth which is closer to the reconstruction than the evaluation threshold distance (*). Larger is better.F1 score [%]: The harmonic mean of accuracy and completeness, used to rank methods based on both metrics. Larger is better.Time [s]: The runtime of the method. Smaller is better.
(*) For exact definitions, detailing how potentially incomplete ground truth is taken into account, see our paper .
The datasets are grouped into different categories, and result averages are computed for a category and method if results of the method are available for all datasets within the category.
Note that the category "all" includes both the high-res multi-view and the low-res many-view scenarios.
Methods with suffix _ROB may participate in the Robust Vision Challenge .
training datasets , ground truth and accuracy / completeness visualizations are also available. The visualizations may not work with mobile browsers.
Set : Test
Metric : Accuracy [%]
Tolerance : 2cm
Method
Info all high-res indoor outdoor botani. boulde. bridge door exhibi. lectur. living. lounge observ. old co. statue terrac.
LTVRE_ROB 86.10 1 93.04 1 93.44 1 91.82 2 93.73 4 85.72 4 96.79 1 95.51 2 88.86 4 95.63 2 96.99 2 84.81 10 93.16 14 91.10 1 97.53 8 96.58 1 Andreas Kuhn, Heiko Hirschmüller, Daniel Scharstein, Helmut Mayer: A TV Prior for High-Quality Scalable Multi-View Stereo Reconstruction . International Journal of Computer Vision 2016 COLMAP_ROB 83.01 3 91.97 2 91.95 2 92.04 1 93.68 5 86.15 3 93.37 3 95.10 3 90.28 2 92.11 8 95.03 7 89.75 6 94.93 6 79.95 9 98.27 6 95.04 6 Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo . ECCV 2016 ACMH+ 77.89 11 90.73 3 91.50 3 88.44 13 93.00 8 76.26 20 89.65 6 95.55 1 87.72 6 93.89 5 94.94 8 91.62 1 93.44 10 78.35 13 98.76 2 95.61 3 ACMM 79.39 9 90.65 4 90.99 5 89.63 9 95.03 2 83.53 9 89.62 7 93.35 11 79.20 11 93.83 6 93.60 13 90.81 3 92.96 15 85.02 2 98.45 4 92.41 15 Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo . CVPR 2019 LS3D 90.60 5 90.34 8 91.38 3 89.03 20 84.23 8 93.58 2 94.68 4 90.65 1 90.33 11 93.95 11 84.50 11 95.37 3 80.87 7 95.50 19 94.53 9 ACMP 79.16 10 90.54 6 90.60 7 90.35 6 94.36 3 80.98 13 89.16 11 94.05 6 79.03 13 92.87 7 93.74 12 90.60 5 95.21 5 82.87 5 98.76 2 94.87 7 Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo . AAAI 2020 PMVS 80.48 7 90.08 7 90.66 6 88.34 15 96.71 1 92.97 2 90.31 5 89.57 23 88.51 5 90.11 12 95.12 6 90.98 2 88.92 23 77.55 15 97.07 11 83.13 29 Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis . PAMI (2010) 3Dnovator+ 83.92 2 89.97 8 89.66 9 90.92 4 91.27 13 81.66 12 91.25 4 94.13 5 88.89 3 88.83 16 94.26 10 77.55 21 95.86 1 84.88 3 95.87 17 95.24 5 ACMH 76.49 14 89.34 9 91.14 4 83.96 23 92.50 10 70.36 26 89.55 8 93.84 7 81.89 9 94.70 4 95.44 4 90.69 4 88.31 25 83.33 4 98.30 5 93.20 12 Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo . CVPR 2019 AP-MVS 89.13 10 88.73 11 90.33 7 92.75 9 82.40 11 89.43 10 93.54 8 79.07 12 90.66 10 91.57 18 85.52 9 93.30 12 79.00 12 97.05 12 95.29 4 wuykxyi23d 88.46 11 88.80 10 87.44 17 90.96 14 93.03 1 85.85 20 81.96 31 74.58 18 98.58 1 97.29 1 87.73 7 87.31 26 82.84 6 99.41 1 81.99 31 DeepC-MVS 81.61 5 88.02 12 87.72 12 88.94 11 93.65 6 78.82 17 87.42 16 93.51 9 72.38 21 89.70 14 91.79 17 83.54 12 95.39 2 80.35 8 97.12 9 92.60 13 DeepC-MVS_fast 81.74 4 87.68 13 87.42 13 88.47 12 93.22 7 77.58 19 88.54 14 92.61 13 74.49 19 88.75 17 90.91 20 81.96 17 95.37 3 79.93 10 96.33 14 92.47 14 DeepPCF-MVS 81.24 6 87.28 14 86.21 15 90.49 5 91.48 12 84.90 7 83.41 23 92.38 15 70.25 24 89.35 15 90.68 21 82.85 14 94.57 7 79.55 11 95.95 15 92.00 17 TAPA-MVS 77.73 12 85.71 15 84.83 17 88.37 14 88.78 21 79.72 16 87.15 17 93.50 10 69.17 25 85.80 19 89.56 23 80.76 18 92.13 18 73.21 19 95.51 18 93.25 11 Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo . ICCV 2019 F-COLMAP 84.97 16 83.42 20 89.63 9 92.39 11 83.40 10 88.83 12 91.92 16 73.19 20 80.18 28 89.15 24 77.04 22 93.28 13 65.82 24 92.28 24 92.21 16 3Dnovator 80.37 8 84.80 17 84.71 18 85.06 20 86.36 25 74.71 22 88.77 13 90.00 21 75.65 16 84.96 22 93.17 14 74.06 25 91.19 19 78.28 14 91.09 26 89.29 21 Gipuma 84.44 18 86.33 14 78.78 28 84.20 28 73.57 23 89.55 8 90.44 19 84.24 7 84.38 23 94.89 9 76.35 24 80.40 31 76.14 16 96.80 13 82.36 30 S. Galliani, K. Lasinger, K. Schindler: Massively Parallel Multiview Stereopsis by Surface Normal Diffusion . ICCV 2015 AdaColmap 83.66 19 83.69 19 83.57 24 90.05 17 72.26 25 86.29 18 90.00 21 78.19 14 81.65 27 87.16 29 83.40 13 94.24 8 61.69 28 94.76 21 84.21 28 CLD-MVS 83.18 20 82.64 21 84.79 21 89.05 19 67.82 29 77.93 30 92.52 14 68.33 26 85.07 21 81.54 32 82.06 16 92.96 15 69.35 22 97.91 7 93.57 10 Zhaoxin Li, Wangmeng Zuo, Zhaoqi Wang, Lei Zhang: Confidence-based Large-scale Dense Multi-view Stereo . IEEE Transaction on Image Processing, 2020 ANet_high 83.17 21 85.68 16 75.65 31 81.24 31 45.26 35 79.94 29 92.91 12 83.83 8 91.33 9 96.88 3 80.25 19 85.92 29 68.89 23 95.89 16 95.76 2 LF4IMVS 82.75 22 81.93 23 85.19 19 82.08 30 80.15 15 85.53 22 88.76 26 68.01 28 85.58 20 87.75 25 71.80 29 86.85 27 74.02 17 93.87 23 88.58 23 PCF-MVS 74.62 15 82.15 23 80.92 25 85.84 18 89.43 18 72.30 24 80.53 28 91.82 17 57.36 33 87.81 18 89.92 22 77.67 20 93.63 9 58.69 31 95.08 20 91.58 18 Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction . GCPR 2019 PLC 73.85 16 82.09 24 80.31 26 87.45 16 90.86 15 80.29 14 85.88 19 90.65 18 68.17 27 76.32 33 86.33 30 73.12 27 92.61 17 61.40 30 90.02 29 89.44 20 Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency . Pacific Graphics 2019 OpenMVS 76.72 13 81.98 25 82.00 22 81.93 25 84.42 26 68.22 27 88.50 15 89.48 24 66.92 30 81.80 26 91.86 16 72.59 28 90.16 22 71.19 20 91.25 25 87.40 24 MG-MVS 80.32 26 80.94 24 78.47 30 88.18 23 52.62 33 82.29 24 85.01 28 72.01 22 79.24 29 92.54 15 69.36 30 93.36 11 70.65 21 89.19 30 89.45 19 MAR-MVS 80.24 27 78.74 29 84.73 22 86.87 24 78.18 18 85.75 21 87.81 27 65.67 31 77.84 30 78.50 33 73.79 26 90.53 21 61.59 29 90.87 27 85.49 27 Zhenyu Xu, Yiguang Liu, Xuelei Shi, Ying Wang, Yunan Zheng: MARMVS: Matching Ambiguity Reduced Multiple View Stereo for Efficient Large Scale Scene Reconstruction . CVPR 2020 MSDG 80.06 28 79.99 27 80.25 26 83.91 29 68.04 28 77.51 31 89.19 25 77.65 15 81.94 25 83.45 31 76.37 23 86.31 28 63.31 26 86.59 31 86.41 25 OpenMVS_ROB 70.19 17 77.77 29 77.46 30 78.71 29 84.39 27 61.15 31 81.18 26 82.52 29 62.45 32 83.34 24 87.37 28 66.20 31 88.66 24 64.69 25 85.02 32 86.32 26 wuyk23d 75.13 30 79.30 28 62.63 33 75.56 33 75.18 21 80.89 27 73.10 32 75.06 17 94.76 3 95.32 5 87.73 7 52.85 35 34.16 35 97.11 10 59.85 33 CMPMVS 59.41 18 75.12 31 73.57 31 79.77 27 75.84 32 67.22 30 81.21 25 82.18 30 50.78 34 76.50 32 87.66 26 55.20 33 82.99 30 62.17 27 90.64 28 89.09 22 M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces . CVPR 2011 FPMVS 72.29 32 72.00 32 73.14 32 88.63 22 85.00 5 74.65 32 67.39 33 71.94 23 77.80 31 87.66 26 50.48 35 75.83 32 49.95 32 79.51 33 58.58 35 PVSNet 51.08 19 56.10 33 54.97 33 59.48 34 75.12 35 53.28 32 55.16 33 61.89 34 44.30 35 59.16 34 62.48 36 54.22 34 65.91 33 35.40 34 47.01 35 59.25 34 PNet_i23d 52.13 34 51.24 34 54.79 36 75.56 33 45.26 35 54.54 34 52.55 35 66.95 29 57.19 35 65.82 35 13.15 37 63.40 34 36.39 33 39.04 36 55.71 36 MVE 40.22 20 51.82 35 50.47 35 55.87 35 62.66 36 51.91 34 31.61 35 39.28 36 40.65 36 50.76 36 74.98 34 56.24 32 44.67 36 33.94 36 64.11 34 71.04 32 Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment . EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014) DeepMVS_CX 24.13 37 32.95 37 29.49 37 21.63 37 12.07 37 37.95 37 45.07 37 30.84 36 19.21 37 17.94 37 33.06 37 23.69 37 ITE_SJBPF 90.11 8 90.72 16 84.97 6 90.30 20 81.56 10 90.02 13 91.20 19 82.40 15 90.81 20 73.58 18 94.66 22 94.56 8
