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+77.84 459.53 447.18 767.76 470.63 167.68 443.34 751.01 964.98 6
3Dnovator76.31 559.36 548.19 566.81 570.04 267.59 544.27 552.10 762.80 10
TAPA-MVS73.13 958.67 652.34 362.89 1163.68 1063.46 1145.85 458.83 261.53 11
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
PCF-MVS73.52 757.06 748.10 663.03 966.85 762.36 1343.32 852.89 559.86 14
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
ACMP74.13 656.41 844.60 1264.28 661.46 1465.81 737.07 1852.13 665.56 5
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
COLMAP(SR)56.21 944.16 1364.25 763.87 962.07 1441.86 1046.45 1766.80 4
OpenMVScopyleft72.83 1056.18 1045.66 963.19 864.67 865.13 839.60 1451.71 859.76 15
COLMAP(base)55.51 1144.97 1162.54 1360.85 1562.46 1242.24 947.71 1364.29 7
TAPA-MVS(SR)55.07 1243.33 1562.90 1067.41 657.81 2139.94 1346.71 1663.49 8
ACMM73.20 855.01 1343.19 1662.89 1159.60 1766.07 635.89 2050.48 1063.01 9
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
LTVRE_ROB69.57 1253.52 1445.46 1058.89 1558.76 1860.60 1843.91 647.01 1557.32 20
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 1545.79 858.64 1756.31 2560.76 1641.70 1149.88 1158.86 17
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
GSE52.52 1640.81 1960.34 1461.71 1358.23 2038.13 1743.48 2061.07 13
COLMAP_ROBcopyleft66.92 1652.32 1742.45 1758.89 1556.18 2661.09 1538.61 1546.28 1859.41 16
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
ACMH+68.96 1352.02 1843.83 1457.49 2059.85 1651.27 3340.04 1247.62 1461.35 12
LPCS50.60 1940.43 2057.38 2158.37 2056.00 2538.14 1642.71 2257.77 18
IB-MVS68.01 1449.19 2042.00 1853.98 2758.50 1951.98 3035.85 2148.16 1251.46 27
OpenMVS_ROBcopyleft64.09 1848.56 2138.68 2255.15 2461.90 1249.65 3536.02 1941.33 2453.89 24
ACMH67.68 1547.97 2238.24 2454.45 2656.88 2250.08 3434.49 2341.99 2356.40 21
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
CIDER47.21 2338.36 2353.11 2855.37 2855.10 2833.25 2643.46 2148.87 32
Qingshan Xu and Wenbing Tao: Learning Inverse Depth Regression for Multi-View Stereo with Correlation Cost Volume. AAAI 2020
PVSNet_057.27 1946.73 2434.96 2754.57 2555.50 2751.93 3131.22 2738.71 2656.27 22
AttMVS45.85 2539.26 2150.25 3449.36 3551.75 3234.83 2243.70 1949.63 30
PVSNet64.34 1745.78 2636.40 2652.03 2952.49 3049.20 3734.18 2538.61 2754.41 23
test_112645.72 2728.57 3257.15 2262.33 1159.06 1924.22 3232.91 3350.08 29
CasMVSNet(SR_A)45.13 2825.99 3557.89 1856.32 2464.81 919.76 3832.23 3452.54 26
CasMVSNet(base)44.49 2924.93 3657.53 1956.38 2364.76 1018.64 4031.23 3651.43 28
P-MVSNet44.46 3037.09 2549.37 3649.27 3649.30 3634.35 2439.83 2549.54 31
BP-MVSNet43.22 3132.08 2850.65 3352.86 2946.16 4327.25 2936.92 2852.94 25
Christian Sormann, Patrick Knöbelreiter, Andreas Kuhn, Mattia Rossi, Thomas Pock, Friedrich Fraundorfer: BP-MVSNet: Belief-Propagation-Layers for Multi-View-Stereo. 3DV 2020
Pnet-new-42.61 3222.17 4156.24 2358.21 2153.03 2924.08 3320.25 4857.48 19
A-TVSNet + Gipumacopyleft42.12 3329.09 3150.80 3148.28 3756.53 2329.14 2829.04 3747.59 33
test_120540.46 3429.35 3047.87 3747.88 3856.48 2424.37 3134.32 3039.23 41
ANet39.53 3524.60 3749.48 3549.42 3455.75 2622.08 3427.13 4043.25 36
ANet-0.7539.47 3622.61 4050.71 3249.43 3355.75 2618.21 4127.01 4146.96 34
MVSNet38.33 3726.99 3345.89 3840.49 4357.57 2220.52 3633.47 3139.63 40
Pnet_fast37.64 3817.70 4450.93 3049.72 3260.73 1710.59 5524.81 4342.34 37
R-MVSNet36.87 3929.78 2941.61 4242.00 4246.99 3924.73 3034.83 2935.83 42
CasMVSNet(SR_B)32.16 4026.80 3435.73 4434.57 4446.99 3920.55 3533.05 3225.63 49
test_112431.65 4113.28 5443.90 3944.21 4147.60 3814.80 4311.75 5939.88 39
Pnet-blend++31.51 4215.02 4942.51 4047.41 3946.50 416.99 5923.05 4433.62 43
Pnet-blend31.51 4215.02 4942.51 4047.41 3946.50 416.99 5923.05 4433.62 43
MVSNet_plusplus28.64 4412.93 5539.11 4349.80 3121.55 526.17 6119.69 4945.98 35
MVSCRF28.32 4518.18 4335.09 4532.16 4744.37 4514.66 4421.69 4728.73 46
MVSNet_++25.72 4617.38 4631.27 4633.47 4518.28 572.87 6531.90 3542.06 38
Snet23.85 4717.41 4528.15 4732.72 4624.90 5012.22 5122.60 4626.83 47
A1Net23.74 4823.95 3823.61 4921.43 5319.62 5419.93 3727.96 3829.78 45
CPR_FA23.28 4923.62 3923.05 5019.33 5423.00 5119.34 3927.89 3926.81 48
PMVScopyleft37.38 2121.09 5011.49 6027.48 4824.09 4844.44 4415.98 427.01 6113.92 57
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
F/T MVSNet+Gipuma17.31 5114.39 5119.26 5216.83 5619.67 5314.03 4614.75 5321.27 50
MVSNet + Gipuma16.91 5214.12 5318.77 5516.39 5819.13 5613.98 4714.27 5420.78 51
hgnet16.45 5312.31 5719.21 5321.51 5130.31 4812.80 4811.82 575.81 61
DPSNet16.45 5312.31 5719.21 5321.51 5130.31 4812.80 4811.82 575.81 61
MVEpermissive26.22 2216.26 5516.97 4715.79 5711.75 6019.40 5514.45 4519.48 5016.21 53
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
Pnet-eth16.08 5618.30 4214.59 5821.78 507.07 649.66 5626.95 4214.91 55
firsttry15.88 5712.57 5618.08 5617.80 5517.14 5811.54 5313.61 5519.31 52
example15.69 5810.14 6119.39 5123.73 4930.32 4711.10 549.17 604.13 65
unMVSv113.05 5912.03 5913.73 5914.18 5916.18 5912.13 5211.93 5610.84 58
RMVSNet11.09 6015.49 488.17 649.16 6110.05 6112.57 5018.40 525.29 64
metmvs_fine10.03 6114.31 527.18 656.75 656.98 659.52 5719.10 517.81 59
test_1120copyleft9.53 626.82 6211.33 618.74 629.06 627.57 586.06 6216.20 54
unMVSmet7.42 635.82 638.48 637.60 6311.29 605.84 625.81 636.56 60
CMPMVSbinary51.72 207.38 640.03 6712.27 602.37 6734.46 460.06 670.00 670.00 67
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
Cas-MVS_preliminary7.24 653.51 649.72 627.03 647.46 634.87 632.16 6514.65 56
confMetMVS4.86 663.39 655.85 665.59 666.61 662.94 643.83 645.35 63
FADENet0.17 670.15 660.18 670.31 680.16 670.19 660.10 660.08 66
dnet0.00 680.00 680.00 680.00 690.00 680.00 680.00 670.00 67
UnsupFinetunedMVSNet16.83 56