This table lists the benchmark results for the low-res many-view scenario. The following metrics are evaluated:

(*) 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.

Click a dataset result cell to show a visualization of the reconstruction. For training datasets, ground truth and accuracy / completeness visualizations are also available. The visualizations may not work with mobile browsers.




Method Infoalllow-res
many-view
indooroutdoorlakesidesand boxstorage roomstorage room 2tunnel
sort bysorted bysort bysort bysort bysort bysort bysort bysort by
DeepPCF-MVS98.18 396.68 195.10 197.74 198.30 197.78 392.41 297.79 197.12 1
DeepC-MVS_fast98.59 196.44 294.71 297.59 298.20 297.75 492.08 397.35 496.83 2
DeepC-MVS98.44 296.08 394.14 797.37 397.83 597.46 890.75 997.52 296.81 3
PCF-MVS97.08 1395.68 493.36 1097.23 498.16 396.93 1291.43 895.29 1496.61 4
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
ACMM97.58 595.68 494.49 596.48 896.12 1097.20 1192.04 496.94 796.10 7
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
ACMH+97.24 995.61 693.77 996.83 596.69 897.71 590.70 1096.85 896.09 8
TAPA-MVS97.07 1495.48 794.55 396.10 995.68 1297.30 991.73 697.37 395.32 11
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
PLCcopyleft97.94 495.35 894.28 696.06 1095.26 1596.59 1391.61 796.96 696.34 6
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
ACMH97.28 795.29 993.25 1196.65 696.32 997.93 189.82 1196.68 1095.69 10
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
ACMP97.20 1095.11 1093.01 1396.52 795.68 1297.26 1089.74 1296.27 1196.61 4
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
LTVRE_ROB97.16 1195.10 1194.13 895.75 1195.69 1195.73 1692.87 195.39 1395.83 9
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
A-TVSNet + Gipuma94.96 1294.54 495.24 1495.31 1496.58 1491.82 597.27 593.83 13
COLMAP_ROBcopyleft97.56 694.43 1393.09 1295.33 1394.55 1696.37 1589.45 1396.72 995.07 12
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
3Dnovator97.25 892.47 1487.97 1895.47 1297.93 497.85 286.02 1689.92 1990.64 14
3Dnovator+97.12 1292.17 1587.57 1995.24 1497.49 697.68 685.24 1789.91 2090.54 15
OpenMVScopyleft96.50 1591.77 1686.65 2195.19 1697.38 797.66 784.60 1888.69 2190.53 16
R-MVSNet90.93 1789.80 1591.68 1890.81 1994.15 1987.67 1491.93 1890.07 17
CIDER88.55 1883.13 2492.16 1793.95 1795.61 1783.79 2182.47 2486.92 19
Qingshan Xu and Wenbing Tao: Learning Inverse Depth Regression for Multi-View Stereo with Correlation Cost Volume. AAAI 2020
ANet-0.7587.89 1989.58 1686.76 2084.37 2590.29 2084.51 1994.65 1585.61 21
OpenMVS_ROBcopyleft92.34 1687.46 2081.75 2591.27 1992.55 1895.21 1881.21 2482.29 2586.05 20
P-MVSNet85.35 2191.57 1481.21 2478.76 3282.87 2887.56 1595.58 1282.00 24
ANet85.16 2285.20 2285.14 2184.37 2590.29 2082.00 2388.39 2280.78 26
unMVSv183.89 2384.37 2383.58 2284.00 2985.63 2383.17 2285.56 2381.11 25
A1Net82.90 2489.39 1778.57 3185.19 2263.18 3884.36 2094.42 1687.34 18
RMVSNet80.62 2586.71 2076.57 3284.14 2778.09 3180.22 2593.20 1767.48 37
AttMVS80.32 2678.27 2681.68 2378.98 3186.16 2279.76 2676.79 3279.89 27
hgnet78.82 2776.20 2880.57 2784.50 2382.32 2975.56 2876.84 2974.88 28
DPSNet78.82 2776.20 2880.57 2784.50 2382.32 2975.56 2876.84 2974.88 28
MVSCRF77.21 2974.57 3178.98 3079.29 3083.87 2675.01 3074.12 3373.78 32
MVEpermissive76.82 1776.63 3077.52 2776.03 3470.40 3885.49 2475.68 2779.36 2772.20 34
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
example74.48 3165.59 3480.41 2984.10 2883.16 2762.82 3368.36 3673.98 31
Pnet_fast74.39 3264.28 3681.13 2588.04 2073.18 3560.15 3668.41 3482.18 22
Snet74.39 3264.28 3681.13 2588.04 2073.18 3560.15 3668.41 3482.18 22
MVSNet73.65 3470.01 3376.07 3376.39 3377.59 3261.93 3478.09 2874.23 30
MVSNet + Gipuma72.61 3571.20 3273.54 3575.39 3473.46 3465.60 3276.81 3171.78 35
firsttry72.49 3675.56 3070.44 3768.80 3970.25 3769.31 3181.82 2672.28 33
F/T MVSNet+Gipuma69.80 3764.46 3573.36 3675.05 3573.57 3360.74 3568.18 3771.46 36
PMVScopyleft70.75 1858.19 3843.22 3868.18 3873.06 3784.09 2554.23 3832.21 3847.37 38
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
CMPMVSbinary69.68 1911.01 390.97 3917.70 3910.09 4043.01 391.93 390.00 390.00 39
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
dnet0.00 400.00 400.00 400.00 410.00 400.00 400.00 390.00 39
UnsupFinetunedMVSNet75.05 35