Low-resolution many-view benchmark

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
HY-MVS5.90 15.90 195.90 195.90 205.90 205.90 205.90 215.90 195.90 20
ACMH+729.97 2894.58 44641.24 441063.48 45901.73 45959.70 45639.95 45642.52 441329.02 45
PMVScopyleft863.65 3639.40 38530.00 42712.33 39798.00 44888.00 44701.00 46359.00 34451.00 34
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
IB-MVS872.44 41486.92 471152.62 471709.78 481537.96 491610.93 481146.17 491159.07 481980.45 48
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
ACMH993.80 51055.91 45719.19 451280.39 461101.67 461267.33 46717.78 47720.60 451472.17 46
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
ACMM1311.08 81661.53 491487.23 501777.73 491530.09 481690.76 491475.05 521499.41 502112.35 49
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
ACMP1246.57 71631.36 481293.07 491856.88 501568.97 501868.85 501273.91 511312.23 492132.82 50
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
PVSNet1207.94 62116.05 511643.70 522430.96 522140.83 522235.92 511648.06 541639.33 522916.13 52
PVSNet_01661.03 93323.02 562633.37 563782.80 563505.59 573539.91 562550.07 582716.66 564302.89 56
DeepC-MVS_fast2190.19 123761.23 573050.62 584234.97 583851.12 583561.12 583001.56 613099.67 585292.66 58
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
COLMAP_ROBcopyleft2144.59 113311.60 552554.50 553816.33 573382.00 563555.00 572484.00 572625.00 554512.00 57
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
OpenMVS_ROBcopyleft1936.51 101881.32 501172.27 482354.03 512118.19 512241.51 521205.90 501138.64 472702.38 51
PLCcopyleft2858.82 164457.00 603560.00 615055.00 614457.00 614671.00 603360.00 633760.00 596037.00 61
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
CMPMVSbinary2623.62 154160.94 593037.05 574910.20 604057.20 605253.20 613109.40 622964.70 575420.20 60
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
PCF-MVS4335.85 209288.52 645777.29 6411629.34 659919.03 6610475.60 655496.66 666057.92 6414493.40 67
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
3Dnovator2482.06 142364.20 521597.00 512875.67 542548.00 542712.00 541635.00 531559.00 513367.00 54
OpenMVScopyleft2352.88 132568.40 541727.00 533129.33 552833.00 553048.00 551755.00 551699.00 533507.00 55
3Dnovator+3414.41 174057.00 583365.00 594518.33 593909.00 594339.00 592806.00 593924.00 615307.00 59
DeepPCF-MVS3452.42 185328.53 613926.82 626263.00 625101.63 625379.92 634010.62 643843.02 608307.46 62
DeepC-MVS3917.13 195746.11 624462.91 636601.57 635612.31 635379.45 624610.56 654315.25 638812.96 63
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
TAPA-MVS6309.84 2113353.70 6610496.80 6715258.30 6714678.80 6714057.00 6610791.00 6910202.60 6717039.10 68
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
MVEpermissive10021.00 228749.80 633439.50 6012290.00 666275.00 6416332.00 672923.00 603956.00 6214263.00 66
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
LTVRE_ROB20565.47 2313815.20 677669.50 6617912.33 6822283.00 6821104.00 687445.00 687894.00 6610350.00 64
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
tm-dncc10000000.00 8010000000.00 8010000000.00 8110000000.00 8110000000.00 8110000000.00 8210000000.00 8010000000.00 81
tmmvs10000000.00 8010000000.00 8010000000.00 8110000000.00 8110000000.00 8110000000.00 8210000000.00 8010000000.00 81
test_mvsss0.76 50.73 50.78 50.79 50.80 50.74 60.72 50.76 5
MVS_test_10.89 60.90 60.89 60.85 60.90 60.87 70.92 60.93 6
test_MVS0.73 5
test_robustmvs1.42 111.42 111.42 111.42 121.42 11
vp_mvsnet1.10 91.00 71.17 91.10 91.20 90.90 81.10 91.20 9
PVSNet_LR479.50 34374.44 34549.55 35489.44 35511.52 35366.16 35382.72 35647.68 36
CCVNet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
TVSNet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
test31.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
QQQNet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
SVVNet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
ternet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
SGNet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
unsupervisedMVS_cas10000000.00 8010000000.00 8010000000.00 8110000000.00 8110000000.00 8110000000.00 8210000000.00 8010000000.00 81
mvs_zhu_10300.53 40.53 40.53 40.53 40.53 40.53 40.53 40.53 4
PSD-MVSNet1.42 111.42 111.42 111.42 111.42 111.42 121.42 111.42 11
test_120531740.00 6824750.00 6836400.00 6932520.00 6933720.00 6924240.00 7025260.00 6842960.00 69
test_112631740.00 6824750.00 6836400.00 6932520.00 6933720.00 6924240.00 7025260.00 6842960.00 69
test_112431740.00 6824750.00 6836400.00 6932520.00 6933720.00 6924240.00 7025260.00 6842960.00 69
test_1120copyleft31740.00 6824750.00 6836400.00 6932520.00 6933720.00 6924240.00 7025260.00 6842960.00 69
Cas-MVS_preliminary283.71 30277.25 30288.02 31287.28 31287.14 31275.09 32279.41 30289.65 31
BP-MVSNet1.30 101.30 101.30 101.30 101.30 101.30 111.30 101.30 10
Christian Sormann, Patrick Knöbelreiter, Andreas Kuhn, Mattia Rossi, Thomas Pock, Friedrich Fraundorfer: BP-MVSNet: Belief-Propagation-Layers for Multi-View-Stereo. 3DV 2020
MVSNet_++10.85 259.79 2511.56 2610.23 2614.23 269.34 2710.23 2510.23 26
MVSNet_plusplus10.85 259.79 2511.56 2610.23 2614.23 269.34 2710.23 2510.23 26
confMetMVS86.40 2795.00 2780.67 2867.00 28112.00 2893.00 2997.00 2763.00 28
metmvs_fine86.40 2795.00 2780.67 2867.00 28112.00 2893.00 2997.00 2763.00 28
unMVSmet86.40 2795.00 2780.67 2867.00 28112.00 2893.00 2997.00 2763.00 28
Pnet-blend++9.00 209.00 209.00 219.00 219.00 219.00 229.00 209.00 21
Pnet-eth9.00 209.00 209.00 219.00 219.00 219.00 229.00 209.00 21
CasMVSNet(SR_A)1000000.00 751000000.00 751000000.00 761000000.00 761000000.00 761000000.00 771000000.00 751000000.00 76
CasMVSNet(SR_B)1000000.00 751000000.00 751000000.00 761000000.00 761000000.00 761000000.00 771000000.00 751000000.00 76
TAPA-MVS(SR)1338.82 46969.96 461584.73 471327.62 471473.80 47922.00 481017.91 461952.77 47
CasMVSNet(base)1000000.00 751000000.00 751000000.00 761000000.00 761000000.00 761000000.00 771000000.00 751000000.00 76
Pnet-new-9.00 209.00 209.00 219.00 219.00 219.00 229.00 209.00 21
GSE84345.00 7283012.00 7285233.67 7388513.00 7378675.00 7382336.00 7483688.00 7288513.00 73
CPR_FA300.30 31300.30 31300.30 32300.30 32300.30 32300.30 33300.30 31300.30 32
FADENet300.30 31300.30 31300.30 32300.30 32300.30 32300.30 33300.30 31300.30 32
LPCS84345.00 7283012.00 7285233.67 7388513.00 7378675.00 7382336.00 7483688.00 7288513.00 73
COLMAP(SR)1000000.00 751000000.00 751000000.00 761000000.00 761000000.00 761000000.00 771000000.00 751000000.00 76
COLMAP(base)1000000.00 751000000.00 751000000.00 761000000.00 761000000.00 761000000.00 771000000.00 751000000.00 76
Pnet_fast9.00 209.00 209.00 219.00 219.00 219.00 229.00 209.00 21
Snet10000000.00 8010000000.00 8010000000.00 8110000000.00 8110000000.00 8110000000.00 8210000000.00 8010000000.00 81
Pnet-blend9.00 209.00 209.00 219.00 219.00 219.00 229.00 209.00 21
ANet-0.750.21 10.21 10.21 10.21 10.21 10.21 10.21 10.21 1
A1Net0.21 10.21 10.21 10.21 10.21 10.21 10.21 10.21 1
ANet0.21 10.21 10.21 10.21 10.21 10.21 10.21 10.21 1
dnet10000000.00 8010000000.00 8010000000.00 8110000000.00 8110000000.00 8110000000.00 8210000000.00 8010000000.00 81
unMVSv1438.86 33369.40 33485.16 34432.90 34452.10 34398.70 36340.10 33570.50 35
CIDER669.75 42505.36 41779.35 43703.58 42808.57 42513.92 43496.79 36825.90 40
Qingshan Xu and Wenbing Tao: Learning Inverse Depth Regression for Multi-View Stereo with Correlation Cost Volume. AAAI 2020
RMVSNet1.03 71.03 81.03 71.03 71.03 71.03 91.03 71.03 7
firsttry1.03 71.03 81.03 71.03 71.03 71.03 91.03 71.03 7
AttMVS10000000.00 8010000000.00 8010000000.00 8110000000.00 8110000000.00 8110000000.00 8210000000.00 8010000000.00 81
A-TVSNet + Gipumacopyleft9351.57 657502.61 6510584.21 649219.52 6510119.20 647334.37 677670.84 6512413.90 65
MVSCRF279999.00 7499999.00 74399999.00 75999999.00 7599999.00 7599999.00 7699999.00 7499999.00 75
P-MVSNet626.74 37503.44 38708.94 38634.45 36681.77 38490.58 40516.29 42810.60 39
MVSNet + Gipuma622.80 35505.00 39701.33 36642.00 37671.00 36502.00 41508.00 37791.00 37
F/T MVSNet+Gipuma622.80 35505.00 39701.33 36642.00 37671.00 36502.00 41508.00 37791.00 37
UnsupFinetunedMVSNet10000000.00 81
hgnet660.88 39500.81 35767.59 40683.99 39726.46 39489.58 37512.04 39892.32 41
example660.88 39500.81 35767.59 40683.99 39726.46 39489.58 37512.04 39892.32 41
DPSNet660.88 39500.81 35767.59 40683.99 39726.46 39489.58 37512.04 39892.32 41
R-MVSNet2412.80 531890.00 542761.33 532391.00 532612.00 531879.00 561901.00 543281.00 53
MVSNet769.20 43609.10 43875.93 44786.50 43836.20 43593.70 44624.50 431005.10 44