High-res multi-view results

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.

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	all	high-res multi-view	indoor	outdoor	botani.	boulde.	bridge	door	exhibi.	lectur.	living.	lounge	observ.	old co.	statue	terrac.

SteuartSystems-ACMMP		94.94 1	95.02 1	94.69 6	97.22 14	86.17 8	96.92 6	97.20 6	90.90 3	95.04 3	97.27 3	87.92 2	99.34 4	93.93 2	98.80 2	98.57 3
Skip Steuart: Steuart Systems R&D Blog.
NCCC		94.81 2	94.69 3	95.17 2	97.83 8	87.46 2	95.66 19	96.93 12	92.34 2	93.94 4	96.58 10	87.74 4	99.44 2	92.83 3	98.40 8	98.62 1

CP-MVS		94.34 3	94.21 4	94.74 5	98.39 1	86.64 5	97.60 1	97.24 4	88.53 16	92.73 7	97.23 4	85.20 10	99.32 6	92.15 6	98.83 1	98.25 7

DNCC-MVS		94.24 4	94.07 5	94.75 4	98.06 6	86.90 3	95.88 16	96.94 11	85.68 30	95.05 2	97.18 5	87.31 5	99.07 15	91.90 8	98.61 6	98.28 6

DeepPCF-MVS	89.96 1	94.20 5	94.77 2	92.49 31	96.52 21	80.00 67	94.00 48	97.08 8	90.05 6	95.65 1	97.29 2	89.66 1	98.97 21	93.95 1	98.71 4	98.50 4

DeepC-MVS_fast	89.43 2	94.04 6	93.79 6	94.80 3	97.48 11	86.78 4	95.65 20	96.89 13	89.40 9	92.81 6	96.97 7	85.37 9	99.24 9	90.87 11	98.69 5	98.38 5
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
PHI-MVS		93.89 7	93.65 7	94.62 7	96.84 18	86.43 6	96.69 10	97.49 1	85.15 34	93.56 5	96.28 14	85.60 8	99.31 7	92.45 4	98.79 3	98.12 9

MVS_111021_HR		93.45 8	93.31 8	93.84 14	96.99 16	84.84 15	93.24 57	97.24 4	88.76 15	91.60 9	95.85 20	86.07 7	98.66 28	91.91 7	98.16 11	98.03 13

DELS-MVS		93.43 9	93.25 9	93.97 12	95.42 38	85.04 14	93.06 60	97.13 7	90.74 5	91.84 8	95.09 33	86.32 6	99.21 10	91.22 10	98.45 7	97.65 17

DeepC-MVS	88.79 3	93.31 10	92.99 11	94.26 10	96.07 28	85.83 10	94.89 30	96.99 10	89.02 12	89.56 18	97.37 1	82.51 12	99.38 3	92.20 5	98.30 9	97.57 20
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
ACMMP		93.24 11	92.88 12	94.30 9	98.09 5	85.33 12	96.86 7	97.45 2	88.33 17	90.15 14	97.03 6	81.44 17	99.51 1	90.85 12	95.74 30	98.04 12
Qingshan Xu, Weihang Kong, Wenbing Tao, Marc Pollefeys: Multi-Scale Geometric Consistency Guided and Planar Prior Assisted Multi-View Stereo. IEEE Transactions on Pattern Analysis and Machine Intelligence
CSCG		93.23 12	93.05 10	93.76 15	98.04 7	84.07 22	96.22 12	97.37 3	84.15 37	90.05 15	95.66 24	87.77 3	99.15 14	89.91 13	98.27 10	98.07 10

MVS_111021_LR		92.47 13	92.29 13	92.98 21	95.99 30	84.43 20	93.08 59	96.09 29	88.20 18	91.12 10	95.72 23	81.33 18	97.76 52	91.74 9	97.37 16	96.75 30

3Dnovator+	87.14 4	92.42 14	91.37 17	95.55 1	95.63 36	88.73 1	97.07 4	96.77 16	90.84 4	84.02 73	96.62 9	75.95 42	99.34 4	87.77 17	97.68 14	98.59 2

CPTT-MVS		91.99 15	91.80 14	92.55 28	98.24 3	81.98 43	96.76 9	96.49 20	81.89 62	90.24 13	96.44 13	78.59 31	98.61 30	89.68 14	97.85 13	97.06 24

DP-MVS Recon		91.95 16	91.28 18	93.96 13	98.33 2	85.92 9	94.66 36	96.66 18	82.69 53	90.03 16	95.82 21	82.30 13	99.03 17	84.57 33	96.48 26	96.91 27

EPNet		91.79 17	91.02 21	94.10 11	90.10 104	85.25 13	96.03 15	92.05 88	92.83 1	87.39 34	95.78 22	79.39 26	99.01 19	88.13 15	97.48 15	98.05 11

MG-MVS		91.77 18	91.70 15	92.00 41	97.08 15	80.03 66	93.60 53	95.18 47	87.85 20	90.89 11	96.47 12	82.06 15	98.36 36	85.07 30	97.04 17	97.62 18

Vis-MVSNet		91.75 19	91.23 19	93.29 17	95.32 39	83.78 24	96.14 13	95.98 32	89.89 7	90.45 12	96.58 10	75.09 46	98.31 39	84.75 32	96.90 18	97.78 16
Jingyang Zhang, Yao Yao, Shiwei Li, Zixin Luo, Tian Fang: Visibility-aware Multiview Stereo Network. BMVC 2020
3Dnovator	86.66 5	91.73 20	90.82 23	94.44 8	94.59 53	86.37 7	97.18 3	97.02 9	89.20 11	84.31 72	96.66 8	73.74 53	99.17 12	86.74 25	97.96 12	97.79 15

EPP-MVSNet		91.70 21	91.56 16	92.13 40	95.88 31	80.50 63	97.33 2	95.25 43	86.15 28	89.76 17	95.60 25	83.42 11	98.32 38	87.37 21	93.25 47	97.56 21

IS-MVSNet		91.43 22	91.09 20	92.46 32	95.87 33	81.38 49	96.95 5	93.69 71	89.72 8	89.50 20	95.98 18	78.57 32	97.77 51	83.02 36	96.50 25	98.22 8

PVSNet_Blended_VisFu		91.38 23	90.91 22	92.80 24	96.39 22	83.17 30	94.87 31	96.66 18	83.29 44	89.27 21	94.46 39	80.29 22	99.17 12	87.57 20	95.37 32	96.05 41

OMC-MVS		91.23 24	90.62 25	93.08 18	96.27 25	84.07 22	93.52 54	95.93 33	86.95 22	89.51 19	96.13 16	78.50 33	98.35 37	85.84 28	92.90 51	96.83 29

PAPM_NR		91.22 25	90.78 24	92.52 30	97.60 10	81.46 48	94.37 42	96.24 23	86.39 27	87.41 33	94.80 36	82.06 15	98.48 32	82.80 37	95.37 32	97.61 19

PVSNet_Blended		90.73 26	90.32 26	91.98 42	96.12 26	81.25 51	92.55 67	96.83 14	82.04 59	89.10 23	92.56 66	81.04 19	98.85 23	86.72 26	95.91 29	95.84 45

API-MVS		90.66 27	90.07 27	92.45 33	96.36 23	84.57 18	96.06 14	95.22 46	82.39 54	89.13 22	94.27 44	80.32 21	98.46 33	80.16 49	96.71 21	94.33 69

MAR-MVS		90.30 28	89.37 30	93.07 20	96.61 20	84.48 19	95.68 18	95.67 38	82.36 55	87.85 28	92.85 61	76.63 41	98.80 25	80.01 50	96.68 22	95.91 42
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
PAPR		90.02 29	89.27 31	92.29 36	95.78 34	80.95 56	92.68 66	96.22 24	81.91 61	86.66 38	93.75 54	82.23 14	98.44 35	79.40 54	94.79 36	97.48 22

PVSNet_BlendedMVS		89.98 30	89.70 28	90.82 62	96.12 26	81.25 51	93.92 49	96.83 14	83.49 43	89.10 23	92.26 71	81.04 19	98.85 23	86.72 26	87.86 79	92.35 96

UGNet		89.95 31	88.95 33	92.95 22	94.51 54	83.31 28	95.70 17	95.23 44	89.37 10	87.58 32	93.94 48	64.00 94	98.78 26	83.92 34	96.31 27	96.74 31
Wanjuan Su, Qingshan Xu, Wenbing Tao: Uncertainty-guided Multi-view Stereo Network for Depth Estimation. IEEE Transactions on Circuits and Systems for Video Technology, 2022
AdaColmap		89.89 32	89.07 32	92.37 34	97.41 12	83.03 34	94.42 38	95.92 34	82.81 52	86.34 42	94.65 37	73.89 51	99.02 18	80.69 47	95.51 31	95.05 58

WTY-MVS		89.60 33	88.92 34	91.67 45	95.47 37	81.15 54	92.38 70	94.78 59	83.11 47	89.06 25	94.32 42	78.67 30	96.61 76	81.57 44	90.89 56	97.24 23

Vis-MVSNet (Re-imp)		89.59 34	89.44 29	90.03 72	95.74 35	75.85 89	95.61 21	90.80 102	87.66 21	87.83 29	95.40 27	76.79 38	96.46 80	78.37 56	96.73 20	97.80 14

QAPM		89.51 35	88.15 45	93.59 16	94.92 45	84.58 17	96.82 8	96.70 17	78.43 81	83.41 81	96.19 15	73.18 55	99.30 8	77.11 68	96.54 24	96.89 28

114514_t		89.51 35	88.50 38	92.54 29	98.11 4	81.99 42	95.16 25	96.36 22	70.19 114	85.81 46	95.25 28	76.70 39	98.63 29	82.07 42	96.86 19	97.00 25

CLD-MVS		89.47 37	88.90 35	91.18 51	94.22 60	82.07 41	92.13 74	96.09 29	87.90 19	85.37 56	92.45 67	74.38 48	97.56 56	87.15 22	90.43 58	93.93 75
Zhaoxin Li, Wangmeng Zuo, Zhaoqi Wang, Lei Zhang: Confidence-based Large-scale Dense Multi-view Stereo. IEEE Transaction on Image Processing, 2020
CDS-MVSNet		89.45 38	88.51 37	92.29 36	93.62 72	83.61 27	93.01 61	94.68 61	81.95 60	87.82 30	93.24 58	78.69 29	96.99 66	80.34 48	93.23 48	96.28 35
Khang Truong Giang, Soohwan Song, Sungho Jo: Curvature-guided dynamic scale networks for Multi-view Stereo. ICLR 2022
LPG-MVS_test		89.45 38	88.90 35	91.12 52	94.47 55	81.49 46	95.30 23	96.14 26	86.73 23	85.45 51	95.16 30	69.89 71	98.10 42	87.70 18	89.23 66	93.77 82

ab-mvs		89.41 40	88.35 40	92.60 26	95.15 43	82.65 39	92.20 73	95.60 39	83.97 38	88.55 27	93.70 55	74.16 50	98.21 41	82.46 40	89.37 62	96.94 26

TAMVS		89.21 41	88.29 42	91.96 43	93.71 71	82.62 40	93.30 55	94.19 66	82.22 56	87.78 31	93.94 48	78.83 27	96.95 68	77.70 62	92.98 50	96.32 34

ACMM	84.12 9	89.14 42	88.48 39	91.12 52	94.65 52	81.22 53	95.31 22	96.12 28	85.31 33	85.92 45	94.34 40	70.19 70	98.06 45	85.65 29	88.86 71	94.08 72
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
CNLPA		89.07 43	87.98 48	92.34 35	96.87 17	84.78 16	94.08 47	93.24 75	81.41 67	84.46 65	95.13 32	75.57 45	96.62 74	77.21 67	93.84 41	95.61 50

PLC	84.53 7	89.06 44	88.03 47	92.15 39	97.27 13	82.69 38	94.29 43	95.44 41	79.71 73	84.01 74	94.18 45	76.68 40	98.75 27	77.28 66	93.41 46	95.02 59
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
HY-MVS	83.01 12	89.03 45	87.94 49	92.29 36	94.86 48	82.77 36	92.08 75	94.49 63	81.52 66	86.93 36	92.79 63	78.32 35	98.23 40	79.93 51	90.55 57	95.88 43

ACMP	84.23 8	89.01 46	88.35 40	90.99 60	94.73 49	81.27 50	95.07 29	95.89 36	86.48 26	83.67 78	94.30 43	69.33 77	97.99 46	87.10 24	88.55 73	93.72 85
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
sss		88.93 47	88.26 44	90.94 61	94.05 63	80.78 59	91.71 78	95.38 42	81.55 65	88.63 26	93.91 50	75.04 47	95.47 101	82.47 39	91.61 53	96.57 32

MVSTER		88.84 48	88.29 42	90.51 65	92.95 75	80.44 64	93.73 51	95.01 49	84.66 36	87.15 35	93.12 60	72.79 57	97.21 64	87.86 16	87.36 81	93.87 77

OpenMVS	83.78 11	88.74 49	87.29 53	93.08 18	92.70 78	85.39 11	96.57 11	96.43 21	78.74 79	80.85 94	96.07 17	69.64 74	99.01 19	78.01 60	96.65 23	94.83 63

BH-untuned		88.60 50	88.13 46	90.01 74	95.24 42	78.50 77	93.29 56	94.15 67	84.75 35	84.46 65	93.40 56	75.76 43	97.40 60	77.59 63	94.52 38	94.12 70

1112_ss		88.42 51	87.33 52	91.72 44	94.92 45	80.98 55	92.97 63	94.54 62	78.16 84	83.82 76	93.88 51	78.78 28	97.91 48	79.45 53	89.41 61	96.26 36

BH-RMVSNet		88.37 52	87.48 51	91.02 59	95.28 40	79.45 70	92.89 64	93.07 77	85.45 32	86.91 37	94.84 35	70.35 68	97.76 52	73.97 80	94.59 37	95.85 44

IterMVS-LS		88.36 53	87.91 50	89.70 78	93.80 68	78.29 80	93.73 51	95.08 48	85.73 29	84.75 61	91.90 75	79.88 24	96.92 69	83.83 35	82.51 106	93.89 76
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Marc Pollefeys: IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo.
TAPA-MVS	84.62 6	88.16 54	87.01 55	91.62 46	96.64 19	80.65 60	94.39 39	96.21 25	76.38 89	86.19 43	95.44 26	79.75 25	98.08 44	62.75 111	95.29 34	96.13 38
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
F-COLMAP		87.95 55	86.80 57	91.40 47	96.35 24	80.88 58	94.73 34	95.45 40	79.65 74	82.04 86	94.61 38	71.13 63	98.50 31	76.24 71	91.05 55	94.80 65

LS3D		87.89 56	86.32 59	92.59 27	96.07 28	82.92 35	95.23 24	94.92 54	75.66 95	82.89 82	95.98 18	72.48 59	99.21 10	68.43 101	95.23 35	95.64 49

PCF-MVS	84.11 10	87.74 57	86.08 61	92.70 25	94.02 64	84.43 20	89.27 91	95.87 37	73.62 102	84.43 67	94.33 41	78.48 34	98.86 22	70.27 93	94.45 39	94.81 64
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
XXY-MVS		87.65 58	86.85 56	90.03 72	92.14 79	80.60 62	93.76 50	95.23 44	82.94 50	84.60 63	94.02 46	74.27 49	95.49 100	81.04 45	83.68 101	94.01 74

Test_1112_low_res		87.65 58	86.51 58	91.08 55	94.94 44	79.28 76	91.77 76	94.30 65	76.04 94	83.51 80	92.37 69	77.86 36	97.73 54	78.69 55	89.13 68	96.22 37

BH-w/o		87.57 60	87.05 54	89.12 80	94.90 47	77.90 83	92.41 69	93.51 72	82.89 51	83.70 77	91.34 82	75.75 44	97.07 65	75.49 72	93.49 43	92.39 94

FMVSNet3		87.40 61	86.11 60	91.30 48	93.79 70	83.64 26	94.20 45	94.81 58	83.89 39	84.37 68	91.87 76	68.45 81	96.56 77	78.23 57	85.36 91	93.70 86

GBi-Net		87.26 62	85.98 62	91.08 55	94.01 65	83.10 31	95.14 26	94.94 50	83.57 40	84.37 68	91.64 77	66.59 85	96.34 84	78.23 57	85.36 91	93.79 79

test1		87.26 62	85.98 62	91.08 55	94.01 65	83.10 31	95.14 26	94.94 50	83.57 40	84.37 68	91.64 77	66.59 85	96.34 84	78.23 57	85.36 91	93.79 79

DP-MVS		87.25 64	85.36 69	92.90 23	97.65 9	83.24 29	94.81 32	92.00 90	74.99 99	81.92 88	95.00 34	72.66 58	99.05 16	66.92 105	92.33 52	96.40 33

FMVSNet2		87.19 65	85.82 65	91.30 48	94.01 65	83.67 25	94.79 33	94.94 50	83.57 40	83.88 75	92.05 73	66.59 85	96.51 79	77.56 64	85.01 94	93.73 84

TR-MVS		86.78 66	85.76 66	89.82 75	94.37 59	78.41 78	92.47 68	92.83 79	81.11 69	86.36 41	92.40 68	68.73 79	97.48 58	73.75 83	89.85 59	93.57 87

PatchMatch-RL		86.77 67	85.54 67	90.47 66	95.88 31	82.71 37	90.54 82	92.31 84	79.82 72	84.32 71	91.57 81	68.77 78	96.39 82	73.16 85	93.48 45	92.32 97

PAPM		86.68 68	85.39 68	90.53 64	93.05 74	79.33 75	89.79 87	94.77 60	78.82 77	81.95 87	93.24 58	76.81 37	97.30 61	66.94 104	93.16 49	94.95 60

EPNet_dtu		86.49 69	85.94 64	88.14 87	90.24 102	72.82 99	94.11 46	92.20 87	86.66 25	79.42 100	92.36 70	73.52 54	95.81 96	71.26 90	93.66 42	95.80 46

cascas		86.43 70	84.98 71	90.80 63	92.10 81	80.92 57	90.24 83	95.91 35	73.10 105	83.57 79	88.39 101	65.15 92	97.46 59	84.90 31	91.43 54	94.03 73

LTVRE_ROB	82.13 13	86.26 71	84.90 74	90.34 68	94.44 58	81.50 45	92.31 72	94.89 55	83.03 48	79.63 99	92.67 64	69.69 73	97.79 50	71.20 91	86.26 87	91.72 101
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
FMVSNet1		85.85 72	84.11 78	91.08 55	92.81 77	83.10 31	95.14 26	94.94 50	81.64 63	82.68 83	91.64 77	59.01 103	96.34 84	75.37 73	83.78 100	93.79 79

PatchmatchNet		85.85 72	84.70 75	89.29 79	91.76 84	75.54 90	88.49 99	91.30 98	81.63 64	85.05 57	88.70 98	71.71 61	96.24 87	74.61 78	89.05 69	96.08 39
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Pablo Speciale, Marc Pollefeys: PatchmatchNet: Learned Multi-View Patchmatch Stereo.
CostFormer		85.77 74	84.94 73	88.26 84	91.16 94	72.58 103	89.47 89	91.04 101	76.26 92	86.45 40	89.97 91	70.74 67	96.86 70	82.35 41	87.07 85	95.34 56

PMMVS		85.71 75	84.96 72	87.95 89	88.90 109	77.09 84	88.68 97	90.06 111	72.32 109	86.47 39	90.76 84	72.15 60	94.40 107	81.78 43	93.49 43	92.36 95

PVSNet	78.82 18	85.55 76	84.65 76	88.23 86	94.72 50	71.93 104	87.12 106	92.75 80	78.80 78	84.95 59	90.53 85	64.43 93	96.71 72	74.74 76	93.86 40	96.06 40

ACMH	80.38 17	85.36 77	83.68 83	90.39 67	94.45 57	80.63 61	94.73 34	94.85 57	82.09 58	77.24 104	92.65 65	60.01 101	97.58 55	72.25 87	84.87 95	92.96 90
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
tpmrst		85.35 78	84.99 70	86.43 104	90.88 96	67.88 113	88.71 96	91.43 97	80.13 71	86.08 44	88.80 96	73.05 56	96.02 92	82.48 38	83.40 103	95.40 53

CR-MVSNet		85.35 78	83.76 82	90.12 70	90.58 98	79.34 73	85.24 110	91.96 91	78.27 82	85.55 49	87.87 107	71.03 64	95.61 97	73.96 81	89.36 63	95.40 53

IB-MVS	80.51 15	85.24 80	83.26 86	91.19 50	92.13 80	79.86 68	91.75 77	91.29 99	83.28 45	80.66 96	88.49 100	61.28 98	98.46 33	80.99 46	79.46 114	95.25 57
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
RPSCF		85.07 81	84.27 77	87.48 93	92.91 76	70.62 109	91.69 79	92.46 82	76.20 93	82.67 84	95.22 29	63.94 95	97.29 62	77.51 65	85.80 89	94.53 67

ACMH+	81.04 14	85.05 82	83.46 85	89.82 75	94.66 51	79.37 71	94.44 37	94.12 68	82.19 57	78.04 102	92.82 62	58.23 105	97.54 57	73.77 82	82.90 105	92.54 91

IterMVS		84.88 83	83.98 80	87.60 91	91.44 90	76.03 88	90.18 84	92.41 83	83.24 46	81.06 92	90.42 86	66.60 84	94.28 108	79.46 52	80.98 110	92.48 92
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Marc Pollefeys: IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo.
MSDG		84.86 84	83.09 87	90.14 69	93.80 68	80.05 65	89.18 94	93.09 76	78.89 76	78.19 101	91.91 74	65.86 91	97.27 63	68.47 100	88.45 75	93.11 89

HyFIR		84.73 85	83.91 81	87.19 99	91.64 87	73.54 95	84.44 113	93.89 69	74.63 101	80.91 93	85.24 111	71.67 62	94.46 105	76.78 69	95.98 28	93.57 87

tpm		84.73 85	84.02 79	86.87 103	90.33 100	68.90 112	89.06 95	89.94 112	80.85 70	85.75 47	89.86 92	68.54 80	95.97 93	77.76 61	84.05 99	95.75 47

tpm2		84.08 87	82.94 88	87.48 93	91.39 91	71.27 105	89.23 93	90.37 105	71.95 111	84.64 62	89.33 93	67.30 83	96.55 78	75.17 75	87.09 84	94.63 66

COLMAP_ROB	80.39 16	83.96 88	82.04 91	89.74 77	95.28 40	79.75 69	94.25 44	92.28 85	75.17 97	78.02 103	93.77 53	58.60 104	97.84 49	65.06 109	85.92 88	91.63 102
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
EPMVS		83.90 89	82.70 89	87.51 92	90.23 103	72.67 101	88.62 98	81.96 123	81.37 68	85.01 58	88.34 102	66.31 88	94.45 106	75.30 74	87.12 83	95.43 52

tpmp4_e23		83.87 90	82.33 90	88.48 82	91.46 89	72.82 99	89.82 86	91.57 94	73.02 107	81.86 89	89.05 94	66.20 89	96.97 67	71.57 89	86.39 86	95.66 48

RPMNet		83.18 91	80.87 95	90.12 70	90.58 98	79.34 73	85.24 110	90.78 103	71.44 113	85.55 49	82.97 115	70.87 66	95.61 97	61.01 112	89.36 63	95.40 53

USDC		82.76 92	81.26 93	87.26 96	91.17 92	74.55 92	89.27 91	93.39 74	78.26 83	75.30 108	92.08 72	54.43 111	96.63 73	71.64 88	85.79 90	90.61 107

Patchmtry		82.71 93	80.93 94	88.06 88	90.05 106	76.37 86	84.74 112	91.96 91	72.28 110	81.32 90	87.87 107	71.03 64	95.50 99	68.97 98	80.15 112	92.32 97

PatchT		82.68 94	81.27 92	86.89 102	90.09 105	70.94 108	84.06 115	90.15 108	74.91 100	85.63 48	83.57 113	69.37 76	94.87 104	65.19 108	88.50 74	94.84 62

MIMVSNet		82.59 95	80.53 96	88.76 81	91.51 88	78.32 79	86.57 108	90.13 109	79.32 75	80.70 95	88.69 99	52.98 112	93.07 113	66.03 107	88.86 71	94.90 61

EG-PatchMatch MVS		82.37 96	80.34 97	88.46 83	90.27 101	79.35 72	92.80 65	94.33 64	77.14 86	73.26 113	90.18 89	47.47 119	96.72 71	70.25 94	87.32 82	89.30 111

tpm cat1		81.96 97	80.27 98	87.01 100	91.09 95	71.02 107	87.38 105	91.53 96	66.25 118	80.17 97	86.35 110	68.22 82	96.15 90	69.16 97	82.29 107	93.86 78

FMVSNet5		81.52 98	79.60 100	87.27 95	91.17 92	77.95 82	91.49 81	92.26 86	76.87 87	76.16 105	87.91 106	51.67 113	92.34 114	67.74 103	81.16 108	91.52 103

dp		81.47 99	80.23 99	85.17 109	89.92 107	65.49 117	86.74 107	90.10 110	76.30 91	81.10 91	87.12 109	62.81 96	95.92 94	68.13 102	79.88 113	94.09 71

test_0402		81.30 100	79.17 102	87.67 90	93.19 73	78.17 81	92.98 62	91.71 93	75.25 96	76.02 106	90.31 87	59.23 102	96.37 83	50.22 121	83.63 102	88.47 115

JIA-IIPM		81.04 101	78.98 104	87.25 97	88.64 110	73.48 96	81.75 119	89.61 114	73.19 104	82.05 85	73.71 122	66.07 90	95.87 95	71.18 92	84.60 97	92.41 93

CMPMVS	59.16 21	80.52 102	79.20 101	84.48 111	83.98 118	67.63 114	89.95 85	93.84 70	64.79 119	66.81 121	91.14 83	57.93 107	95.17 102	76.25 70	88.10 77	90.65 106
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
LF4IMVS		80.37 103	79.07 103	84.27 113	86.64 114	69.87 111	89.39 90	91.05 100	76.38 89	74.97 110	90.00 90	47.85 118	94.25 109	74.55 79	80.82 111	88.69 113

UnsupCasMVSNet_eth		80.07 104	78.27 105	85.46 108	85.24 117	72.63 102	88.45 100	94.87 56	82.99 49	71.64 117	88.07 105	56.34 109	91.75 116	73.48 84	63.36 123	92.01 100

TDRefinement		79.81 105	77.34 107	87.22 98	79.24 122	75.48 91	93.12 58	92.03 89	76.45 88	75.01 109	91.58 80	49.19 116	96.44 81	70.22 96	69.18 121	89.75 109

TinyColmap		79.76 106	77.69 106	85.97 105	91.71 85	73.12 97	89.55 88	90.36 106	75.03 98	72.03 115	90.19 88	46.22 120	96.19 89	63.11 110	81.03 109	88.59 114

OpenMVS_ROB	74.94 19	79.51 107	77.03 109	86.93 101	87.00 113	76.23 87	92.33 71	90.74 104	68.93 116	74.52 111	88.23 103	49.58 115	96.62 74	57.64 117	84.29 98	87.94 117

MIMVSNet1		79.38 108	77.28 108	85.69 106	86.35 116	73.67 94	91.61 80	92.75 80	78.11 85	72.64 114	88.12 104	48.16 117	91.97 115	60.32 113	77.49 115	91.43 105

PVSNet_0	73.20 20	77.22 109	74.83 111	84.37 112	90.70 97	71.10 106	83.09 117	89.67 113	72.81 108	73.93 112	83.13 114	60.79 100	93.70 110	68.54 99	50.84 126	88.30 116

DSMNet-mixed		76.94 110	76.29 110	78.89 114	83.10 120	56.11 123	87.78 102	79.77 125	60.65 122	75.64 107	88.71 97	61.56 97	88.34 120	60.07 115	89.29 65	92.21 99

UnsupCasMVSNet_bld		76.23 111	73.27 112	85.09 110	83.79 119	72.92 98	85.65 109	93.47 73	71.52 112	68.84 120	79.08 119	49.77 114	93.21 111	66.81 106	60.52 124	89.13 112

LP		75.51 112	72.15 114	85.61 107	87.86 112	73.93 93	80.20 121	88.43 115	67.39 117	70.05 118	80.56 117	58.18 106	93.18 112	46.28 123	70.36 120	89.71 110

MVS-HIRNet		73.70 113	72.20 113	78.18 116	91.81 83	56.42 122	82.94 118	82.58 121	55.24 124	68.88 119	66.48 123	55.32 110	95.13 103	58.12 116	88.42 76	83.01 118

new_pmnet		72.15 114	70.13 115	78.20 115	82.95 121	65.68 115	83.91 116	82.40 122	62.94 121	64.47 122	79.82 118	42.85 121	86.26 121	57.41 118	74.44 118	82.65 119

HyFIR lowres test		70.96 115	69.05 116	76.67 117	86.42 115	64.06 118	77.75 123	82.95 120	55.25 123	62.90 123	78.84 120	33.89 123	90.95 118	60.14 114	83.30 104	75.01 123

N_pmnet		68.89 116	68.44 117	70.23 121	89.07 108	28.79 131	88.06 101	19.50 130	69.47 115	71.86 116	84.93 112	61.24 99	91.75 116	54.70 119	77.15 116	90.15 108

FPMVS		64.63 117	62.55 118	70.88 120	70.80 124	56.71 120	84.42 114	84.42 119	51.78 125	49.57 124	81.61 116	23.49 126	81.48 124	40.61 126	76.25 117	74.46 124

no-one		61.56 118	56.58 119	76.49 118	67.80 127	62.76 119	78.13 122	86.11 116	63.16 120	43.24 126	64.70 125	26.12 125	88.95 119	50.84 120	29.15 128	77.77 121

ANet_high		58.88 119	54.22 121	72.86 119	56.50 130	56.67 121	80.75 120	86.00 117	73.09 106	37.39 128	64.63 126	22.17 127	79.49 126	43.51 124	23.96 131	82.43 120

Gipuma		57.99 120	54.91 120	67.24 123	88.51 111	65.59 116	52.21 128	90.33 107	43.58 128	42.84 127	51.18 129	20.29 128	85.07 122	34.77 128	70.45 119	51.05 129
S. Galliani, K. Lasinger, K. Schindler: Massively Parallel Multiview Stereopsis by Surface Normal Diffusion. ICCV 2015
PMVS	47.18 22	52.22 121	48.46 122	63.48 124	45.72 131	46.20 128	73.41 124	78.31 126	41.03 129	30.06 130	65.68 124	6.05 130	83.43 123	30.04 129	65.86 122	60.80 126
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
wuykxyi23d		50.55 122	44.13 124	69.81 122	56.77 128	54.58 125	73.22 125	80.78 124	39.79 130	22.08 132	46.69 130	4.03 132	79.71 125	47.65 122	26.13 130	75.14 122

PNet_i23d		50.48 123	47.18 123	60.36 125	68.59 125	44.56 129	72.75 126	72.61 127	43.92 127	33.91 129	60.19 127	6.16 129	73.52 127	38.50 127	28.04 129	63.01 125

MVE	39.65 23	43.39 124	38.59 125	57.77 126	56.52 129	48.77 127	55.38 127	58.64 128	29.33 131	28.96 131	52.65 128	4.68 131	64.62 128	28.11 130	33.07 127	59.93 127
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
wuyk23d		21.27 125	20.48 126	23.63 128	68.59 125	36.41 130	49.57 129	6.85 131	9.37 132	7.89 133	4.46 131	4.03 132	31.37 130	17.47 131	16.07 132	3.12 130

ab-mvs-re		7.82 126	10.43 127	0.00 129	0.00 132	0.00 132	0.00 130	0.00 132	0.00 133	0.00 134	93.88 51	0.00 134	0.00 131	0.00 132	0.00 133	0.00 131

door								85.33 118

HQP-MVS												37.31 122

ACMP_Plane					94.17 61		94.39 39		88.81 13	85.43 53

HQP5-MVS						81.56 44

HQP-NCC					94.17 61		94.39 39		88.81 13	85.43 53

MDTV_nov1_ep13			83.56 84		91.69 86	69.93 110	87.75 103	91.54 95	78.60 80	84.86 60	88.90 95	69.54 75	96.03 91	70.25 94	88.93 70

ACMMP++															88.01 78

BP-MVS														87.11 23

HQP4-MVS										85.43 53			97.96 47			94.51 68

HQP3-MVS								96.04 31							89.77 60

HQP2-MVS												73.83 52

LGP-MVS_train				91.12 52	94.47 55	81.49 46		96.14 26	86.73 23	85.45 51	95.16 30	69.89 71	98.10 42	87.70 18	89.23 66	93.77 82

NP-MVS											93.98 47

DeepMVS_CX				56.31 127	74.23 123	51.81 126		56.67 129	44.85 126	48.54 125	75.16 121	27.87 124	58.74 129	40.92 125	52.22 125	58.39 128

MDTV_nov1_ep13_2view						55.91 124	87.62 104		73.32 103	84.59 64		70.33 69		74.65 77		95.50 51

ACMMP++_ref															87.47 80

Test By Simon												80.02 23

ITE_SJBPF				88.24 85	91.88 82	77.05 85		92.92 78	85.54 31	80.13 98	93.30 57	57.29 108	96.20 88	72.46 86	84.71 96	91.49 104

High-resolution multi-view benchmark