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-MVS80.84 163.41 153.95 169.71 269.72 471.26 148.73 159.18 168.16 2
DeepC-MVS79.81 262.37 252.77 268.77 368.65 570.30 346.95 258.59 367.37 3
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
DeepC-MVS_fast79.65 362.24 350.96 469.75 169.91 371.08 246.04 355.88 468.27 1
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
3Dnovator+77.84 459.53 447.18 867.76 470.63 167.68 443.34 751.01 1064.98 7
3Dnovator76.31 559.36 548.19 566.81 570.04 267.59 544.27 552.10 862.80 11
tm-dncc58.94 648.02 766.22 666.30 865.64 841.30 1254.73 566.71 5
TAPA-MVS73.13 958.67 752.34 362.89 1363.68 1263.46 1345.85 458.83 261.53 12
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
PCF-MVS73.52 757.06 848.10 663.03 1166.85 762.36 1543.32 852.89 659.86 15
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
ACMP74.13 656.41 944.60 1364.28 761.46 1665.81 737.07 2052.13 765.56 6
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
COLMAP(SR)56.21 1044.16 1464.25 863.87 1162.07 1641.86 1046.45 1866.80 4
OpenMVScopyleft72.83 1056.18 1145.66 1063.19 1064.67 1065.13 1039.60 1551.71 959.76 16
COLMAP(base)55.51 1244.97 1262.54 1560.85 1762.46 1442.24 947.71 1464.29 8
TAPA-MVS(SR)55.07 1343.33 1662.90 1267.41 657.81 2539.94 1446.71 1763.49 9
ACMM73.20 855.01 1443.19 1762.89 1359.60 1966.07 635.89 2350.48 1163.01 10
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
tmmvs53.89 1539.15 2463.73 966.30 865.64 837.77 1940.52 2759.24 18
LTVRE_ROB69.57 1353.52 1645.46 1158.89 1758.76 2160.60 2043.91 647.01 1657.32 22
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
PLCcopyleft70.83 1153.50 1745.79 958.64 1956.31 2860.76 1841.70 1149.88 1258.86 19
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
GSE52.52 1840.81 2060.34 1661.71 1558.23 2338.13 1843.48 2261.07 14
COLMAP_ROBcopyleft66.92 1752.32 1942.45 1858.89 1756.18 2961.09 1738.61 1646.28 1959.41 17
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
ACMH+68.96 1452.02 2043.83 1557.49 2259.85 1851.27 3840.04 1347.62 1561.35 13
LPCS50.60 2140.43 2157.38 2358.37 2356.00 2938.14 1742.71 2457.77 20
HY-MVS69.67 1249.80 2240.39 2256.07 2659.17 2053.68 3336.13 2144.66 2055.36 25
IB-MVS68.01 1549.19 2342.00 1953.98 3158.50 2251.98 3535.85 2448.16 1351.46 31
Christian Sormann, Mattia Rossi, Andreas Kuhn and Friedrich Fraundorfer: IB-MVS: An Iterative Algorithm for Deep Multi-View Stereo based on Binary Decisions. BMVC 2021
OpenMVS_ROBcopyleft64.09 1948.56 2438.68 2555.15 2761.90 1449.65 4036.02 2241.33 2653.89 27
ACMH67.68 1647.97 2538.24 2754.45 3056.88 2550.08 3934.49 2641.99 2556.40 23
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
CIDER47.21 2638.36 2653.11 3355.37 3155.10 3233.25 2943.46 2348.87 37
Qingshan Xu and Wenbing Tao: Learning Inverse Depth Regression for Multi-View Stereo with Correlation Cost Volume. AAAI 2020
PVSNet_057.27 2046.73 2734.96 3054.57 2955.50 3051.93 3631.22 3038.71 2956.27 24
AttMVS45.85 2839.26 2350.25 3949.36 4051.75 3734.83 2543.70 2149.63 35
PVSNet64.34 1845.78 2936.40 2952.03 3452.49 3449.20 4234.18 2838.61 3054.41 26
test_112645.72 3028.57 3657.15 2462.33 1359.06 2124.22 3532.91 3850.08 33
CasMVSNet(SR_A)45.13 3125.99 4057.89 2056.32 2764.81 1119.76 4332.23 3952.54 30
PVSNet_LR44.83 3230.05 3254.69 2853.21 3257.88 2421.84 3838.25 3152.98 28
CasMVSNet(base)44.49 3324.93 4157.53 2156.38 2664.76 1218.64 4531.23 4151.43 32
P-MVSNet44.46 3437.09 2849.37 4149.27 4149.30 4134.35 2739.83 2849.54 36
BP-MVSNet43.22 3532.08 3150.65 3852.86 3346.16 4927.25 3236.92 3252.94 29
Christian Sormann, Patrick Knöbelreiter, Andreas Kuhn, Mattia Rossi, Thomas Pock, Friedrich Fraundorfer: BP-MVSNet: Belief-Propagation-Layers for Multi-View-Stereo. 3DV 2020
mvs_zhu_103042.98 3627.14 3753.53 3252.24 3558.28 2220.89 3933.39 3650.07 34
Pnet-new-42.61 3722.17 4656.24 2558.21 2453.03 3424.08 3620.25 5757.48 21
A-TVSNet + Gipumacopyleft42.12 3829.09 3550.80 3648.28 4256.53 2729.14 3129.04 4247.59 38
test_120540.46 3929.35 3447.87 4247.88 4356.48 2824.37 3434.32 3439.23 46
ANet39.53 4024.60 4249.48 4049.42 3955.75 3022.08 3727.13 4643.25 41
ANet-0.7539.47 4122.61 4550.71 3749.43 3855.75 3018.21 4627.01 4746.96 39
MVSNet38.33 4226.99 3845.89 4340.49 4857.57 2620.52 4133.47 3539.63 45
Pnet_fast37.64 4317.70 5250.93 3549.72 3760.73 1910.59 6924.81 5042.34 42
R-MVSNet36.87 4429.78 3341.61 4742.00 4746.99 4424.73 3334.83 3335.83 47
CasMVSNet(SR_B)32.16 4526.80 3935.73 5034.57 5146.99 4420.55 4033.05 3725.63 56
test_112431.65 4613.28 6843.90 4444.21 4647.60 4314.80 5011.75 7539.88 44
Pnet-blend++31.51 4715.02 6142.51 4547.41 4446.50 476.99 7623.05 5133.62 49
Pnet-blend31.51 4715.02 6142.51 4547.41 4446.50 476.99 7623.05 5133.62 49
unsupervisedMVS_cas30.64 4920.24 4737.57 4935.93 5042.74 5215.36 4825.12 4934.04 48
MVSNet_plusplus28.64 5012.93 6939.11 4849.80 3621.55 696.17 7919.69 5845.98 40
MVSCRF28.32 5118.18 5035.09 5232.16 5444.37 5114.66 5121.69 5428.73 52
MVSNet_++25.72 5217.38 5431.27 5333.47 5218.28 742.87 8431.90 4042.06 43
MVS_test_125.14 5318.49 4829.57 5437.50 4927.99 599.50 7227.48 4523.21 63
vp_mvsnet23.86 546.73 7935.28 5131.55 5546.94 466.47 786.99 7827.34 53
Snet23.85 5517.41 5328.15 5532.72 5324.90 6012.22 6222.60 5326.83 54
A1Net23.74 5623.95 4323.61 5921.43 6219.62 7119.93 4227.96 4329.78 51
CPR_FA23.28 5723.62 4423.05 6019.33 6423.00 6419.34 4427.89 4426.81 55
PMVScopyleft37.38 2221.09 5811.49 7627.48 5624.09 5744.44 5015.98 477.01 7713.92 74
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
test320.33 5917.79 5122.02 6117.88 6624.31 6114.91 4920.68 5623.87 57
TVSNet20.06 6017.25 5521.93 6218.25 6523.77 6213.74 5620.77 5523.78 58
test_mvsss19.92 6111.93 7525.25 5724.18 5632.22 547.42 7516.44 6719.34 68
CCVNet19.37 6212.14 7324.19 5820.50 6328.61 5811.03 6813.25 7123.48 59
QQQNet18.88 6315.19 6021.34 6317.26 6823.30 6313.79 5516.60 6423.48 59
SGNet18.50 6416.17 5720.06 6616.84 6922.87 6513.42 5718.92 6120.47 66
SVVNet18.01 6514.28 6520.49 6415.75 7422.30 6611.95 6416.60 6423.43 61
ternet18.01 6514.28 6520.49 6415.75 7422.30 6611.95 6416.60 6423.43 61
PSD-MVSNet17.93 6715.59 5819.48 6716.64 7222.17 6813.02 5818.15 6319.64 67
F/T MVSNet+Gipuma17.31 6814.39 6319.26 6916.83 7019.67 7014.03 5314.75 6821.27 64
MVSNet + Gipuma16.91 6914.12 6718.77 7216.39 7319.13 7313.98 5414.27 6920.78 65
hgnet16.45 7012.31 7119.21 7021.51 6030.31 5612.80 5911.82 735.81 78
DPSNet16.45 7012.31 7119.21 7021.51 6030.31 5612.80 5911.82 735.81 78
MVEpermissive26.22 2316.26 7216.97 5615.79 7411.75 7719.40 7214.45 5219.48 5916.21 70
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
Pnet-eth16.08 7318.30 4914.59 7521.78 597.07 819.66 7026.95 4814.91 72
firsttry15.88 7412.57 7018.08 7317.80 6717.14 7511.54 6613.61 7019.31 69
example15.69 7510.14 7719.39 6823.73 5830.32 5511.10 679.17 764.13 82
unMVSv113.05 7612.03 7413.73 7614.18 7616.18 7612.13 6311.93 7210.84 75
RMVSNet11.09 7715.49 598.17 819.16 7810.05 7812.57 6118.40 625.29 81
metmvs_fine10.03 7814.31 647.18 826.75 826.98 829.52 7119.10 607.81 76
test_1120copyleft9.53 796.82 7811.33 788.74 799.06 797.57 736.06 7916.20 71
unMVSmet7.42 805.82 808.48 807.60 8011.29 775.84 805.81 806.56 77
CMPMVSbinary51.72 217.38 810.03 8412.27 772.37 8534.46 530.06 860.00 840.00 85
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
Cas-MVS_preliminary7.24 823.51 819.72 797.03 817.46 804.87 822.16 8214.65 73
confMetMVS4.86 833.39 825.85 835.59 836.61 832.94 833.83 815.35 80
FADENet0.17 840.15 830.18 850.31 860.16 850.19 850.10 830.08 84
dnet0.00 850.00 850.00 860.00 870.00 860.00 870.00 840.00 85
test_MVS7.47 74
test_robustmvs3.18 845.44 842.71 844.94 811.39 83
UnsupFinetunedMVSNet16.83 70