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.

DeepC-MVS_fast	93.89 2	96.93 1	96.64 2	97.78 2	98.64 6	94.30 2	97.41 20	98.04 16	94.81 2	96.59 3	98.37 4	91.24 5	99.64 11	95.16 6	99.52 2	99.42 3
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
PHI-MVS		96.77 2	96.46 4	97.71 3	98.40 12	94.07 4	98.21 10	98.45 5	89.86 28	97.11 2	98.01 10	92.52 3	99.69 9	96.03 3	99.53 1	99.36 4

MVS_111021_HR		96.68 3	96.58 3	96.99 9	98.46 9	92.31 13	96.20 48	98.90 1	94.30 5	95.86 4	97.74 16	92.33 4	99.38 23	96.04 2	99.42 4	99.28 6

DeepPCF-MVS	93.97 1	96.61 4	97.09 1	95.15 42	98.09 20	86.63 64	96.00 51	98.15 10	95.43 1	97.95 1	98.56 2	93.40 1	99.36 25	96.77 1	99.48 3	99.45 1

ACMMP		96.27 5	95.93 6	97.28 7	99.24 1	92.62 9	98.25 9	98.81 2	92.99 12	94.56 12	98.39 3	88.96 12	99.85 1	94.57 8	97.63 26	99.36 4

MVS_111021_LR		96.24 6	96.19 5	96.39 22	98.23 18	91.35 26	96.24 47	98.79 3	93.99 7	95.80 5	97.65 18	89.92 7	99.24 29	95.87 4	99.20 8	98.58 20

DeepC-MVS	93.07 3	96.06 7	95.66 8	97.29 6	97.96 22	93.17 6	97.30 24	98.06 14	93.92 8	93.38 23	98.66 1	86.83 22	99.73 6	95.60 5	99.22 7	98.96 14
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
CSCG		96.05 8	95.91 7	96.46 20	99.24 1	90.47 38	98.30 6	98.57 4	89.01 39	93.97 19	97.57 21	92.62 2	99.76 4	94.66 7	99.27 6	99.15 8

DP-MVS Recon		95.68 9	95.12 10	97.37 5	99.19 3	94.19 3	97.03 28	98.08 13	88.35 48	95.09 7	97.65 18	89.97 6	99.48 19	92.08 16	98.59 15	98.44 28

MG-MVS		95.61 10	95.38 9	96.31 24	98.42 11	90.53 37	96.04 50	97.48 25	93.47 9	95.67 6	98.10 7	89.17 10	99.25 28	91.27 22	98.77 11	99.13 10

3Dnovator+	91.43 4	95.40 11	94.48 18	98.16 1	96.90 40	95.34 1	98.48 3	97.87 17	94.65 3	88.53 60	98.02 9	83.69 38	99.71 8	93.18 13	98.96 9	99.44 2

Vis-MVSNet		95.23 12	94.81 12	96.51 16	97.18 33	91.58 23	98.26 8	98.12 11	94.38 4	94.90 8	98.15 6	82.28 43	98.92 41	91.45 20	98.58 16	99.01 12
Jingyang Zhang, Yao Yao, Shiwei Li, Zixin Luo, Tian Fang: Visibility-aware Multiview Stereo Network. BMVC 2020
EPP-MVSNet		95.22 13	95.04 11	95.76 36	97.49 31	89.56 44	98.67 1	97.00 43	90.69 20	94.24 17	97.62 20	89.79 8	98.81 43	93.39 12	96.49 36	98.92 16

3Dnovator	91.36 5	95.19 14	94.44 19	97.44 4	96.56 45	93.36 5	98.65 2	98.36 6	94.12 6	89.25 55	98.06 8	82.20 44	99.77 3	93.41 11	99.32 5	99.18 7

OMC-MVS		95.09 15	94.70 13	96.25 27	98.46 9	91.28 27	96.43 40	97.57 24	92.04 15	94.77 10	97.96 11	87.01 21	99.09 36	91.31 21	96.77 33	98.36 32

PAPM_NR		95.01 16	94.59 14	96.26 26	98.89 4	90.68 35	97.24 25	97.73 19	91.80 16	92.93 29	96.62 38	89.13 11	99.14 32	89.21 32	97.78 25	98.97 13

IS-MVSNet		94.90 17	94.52 16	96.05 30	97.67 27	90.56 36	98.44 4	96.22 62	93.21 11	93.99 18	97.74 16	85.55 28	98.45 51	89.98 27	97.86 23	99.14 9

PVSNet_Blended		94.87 18	94.56 15	95.81 34	98.27 16	89.46 46	95.47 59	98.36 6	88.84 43	94.36 14	96.09 47	88.02 16	99.58 12	93.44 9	98.18 19	98.40 30

API-MVS		94.84 19	94.49 17	95.90 32	97.90 24	92.00 18	97.80 12	97.48 25	89.19 36	94.81 9	96.71 32	88.84 13	99.17 31	88.91 35	98.76 12	96.53 54

WTY-MVS		94.71 20	94.02 20	96.79 10	97.71 26	92.05 17	96.59 36	97.35 35	90.61 23	94.64 11	96.93 29	86.41 25	99.39 22	91.20 24	94.71 46	98.94 15

sss		94.51 21	93.80 23	96.64 11	97.07 37	91.97 19	96.32 43	98.06 14	88.94 42	94.50 13	96.78 31	84.60 33	99.27 27	91.90 17	96.02 38	98.68 19

AdaColmap		94.34 22	93.68 24	96.31 24	98.59 7	91.68 21	96.59 36	97.81 18	89.87 27	92.15 33	97.06 28	83.62 39	99.54 15	89.34 31	98.07 21	97.70 44

CNLPA		94.28 23	93.53 26	96.52 14	98.38 14	92.55 10	96.59 36	96.88 50	90.13 25	91.91 36	97.24 23	85.21 29	99.09 36	87.64 41	97.83 24	97.92 40

MAR-MVS		94.22 24	93.46 27	96.51 16	98.00 21	92.19 16	97.67 13	97.47 27	88.13 50	93.00 26	95.84 50	84.86 32	99.51 17	87.99 39	98.17 20	97.83 41
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		94.18 25	93.42 30	96.48 18	97.64 28	91.42 25	95.55 56	97.71 21	88.99 40	92.34 32	95.82 51	89.19 9	99.11 34	86.14 50	97.38 30	98.90 17

Vis-MVSNet (Re-imp)		94.15 26	93.88 22	94.95 47	97.61 29	87.92 55	98.10 11	95.80 68	92.22 14	93.02 25	97.45 22	84.53 35	97.91 67	88.24 37	97.97 22	99.02 11

anonymous		94.14 27	93.54 25	95.93 31	96.18 55	91.46 24	96.33 42	97.04 42	88.97 41	93.56 20	96.51 40	87.55 19	97.89 68	89.80 29	95.95 39	98.44 28

PLC	91.00 6	94.11 28	93.43 29	96.13 29	98.58 8	91.15 30	96.69 35	97.39 32	87.29 58	91.37 42	96.71 32	88.39 15	99.52 16	87.33 43	97.13 32	97.73 42
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
PVSNet_BlendedMVS		94.06 29	93.92 21	94.47 53	98.27 16	89.46 46	96.73 33	98.36 6	90.17 24	94.36 14	95.24 58	88.02 16	99.58 12	93.44 9	90.72 63	94.36 72

TAMVS		94.01 30	93.46 27	95.64 37	96.16 57	90.45 39	96.71 34	96.89 49	89.27 35	93.46 22	96.92 30	87.29 20	97.94 66	88.70 36	95.74 41	98.53 21

HY-MVS	89.66 9	93.87 31	92.95 32	96.63 13	97.10 36	92.49 12	95.64 54	96.64 56	89.05 38	93.00 26	95.79 53	85.77 27	99.45 21	89.16 34	94.35 47	97.96 39

F-COLMAP		93.58 32	92.98 31	95.37 40	98.40 12	88.98 50	97.18 27	97.29 36	87.75 54	90.49 44	97.10 27	85.21 29	99.50 18	86.70 46	96.72 34	97.63 45

ab-mvs		93.57 33	92.55 35	96.64 11	97.28 32	91.96 20	95.40 60	97.45 30	89.81 30	93.22 24	96.28 44	79.62 52	99.46 20	90.74 26	93.11 53	98.50 24

LS3D		93.57 33	92.61 34	96.47 19	97.59 30	91.61 22	97.67 13	97.72 20	85.17 65	90.29 45	98.34 5	84.60 33	99.73 6	83.85 62	98.27 18	98.06 37

UGNet		93.55 35	92.80 33	95.79 35	96.63 43	89.98 40	97.57 16	97.47 27	93.38 10	92.11 34	95.94 48	72.50 69	99.38 23	90.94 25	98.65 13	98.01 38

1112_ss		93.37 36	92.42 38	96.21 28	97.05 38	90.99 31	96.31 44	96.72 53	86.87 59	89.83 48	96.69 34	86.51 24	99.14 32	88.12 38	93.67 49	98.50 24

QAPM		93.13 37	91.83 45	97.03 8	96.71 42	92.80 7	98.34 5	98.19 9	81.23 79	88.28 62	97.89 12	80.38 48	99.79 2	86.60 47	98.86 10	98.51 23

CLD-MVS		92.98 38	92.53 36	94.32 56	96.12 58	89.20 48	95.28 62	97.47 27	92.66 13	89.90 47	95.62 55	80.58 46	98.40 52	92.73 14	92.40 56	95.38 60
Zhaoxin Li, Wangmeng Zuo, Zhaoqi Wang, Lei Zhang: Confidence-based Large-scale Dense Multi-view Stereo. IEEE Transaction on Image Processing, 2020
ACMM	89.79 8	92.96 39	92.50 37	94.35 55	96.30 52	88.71 52	97.58 15	97.36 34	91.40 19	90.53 43	96.65 36	79.77 50	98.75 45	91.24 23	91.64 59	95.59 58
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
PatchMatch-RL		92.90 40	92.02 41	95.56 38	98.19 19	90.80 34	95.27 63	97.18 39	87.96 51	91.86 38	95.68 54	80.44 47	98.99 39	84.01 60	97.54 28	96.89 51

OpenMVS	89.19 12	92.86 41	91.68 47	96.40 21	95.34 63	92.73 8	98.27 7	98.12 11	84.86 68	85.78 70	97.75 15	78.89 53	99.74 5	87.50 42	98.65 13	96.73 53

PMMVS		92.86 41	92.34 39	94.42 54	94.92 71	86.73 63	94.53 65	96.38 59	84.78 69	94.27 16	95.12 60	83.13 40	98.40 52	91.47 19	96.49 36	98.12 35

Test_1112_low_res		92.84 43	91.84 44	95.85 33	97.04 39	89.97 41	95.53 57	96.64 56	85.38 64	89.65 49	95.18 59	85.86 26	99.10 35	87.70 40	93.58 52	98.49 26

DP-MVS		92.76 44	91.51 48	96.52 14	98.77 5	90.99 31	97.38 22	96.08 64	82.38 77	89.29 54	97.87 13	83.77 37	99.69 9	81.37 72	96.69 35	98.89 18

BH-RMVSNet		92.72 45	91.97 42	94.97 45	97.16 34	87.99 54	96.15 49	95.60 69	90.62 22	91.87 37	97.15 26	78.41 57	98.57 48	83.16 63	97.60 27	98.36 32

ACMP	89.59 10	92.62 46	92.14 40	94.05 59	96.40 50	88.20 53	97.36 23	97.25 38	91.52 17	88.30 61	96.64 37	78.46 56	98.72 46	91.86 18	91.48 60	95.23 66
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
TAPA-MVS	90.10 7	92.30 47	91.22 51	95.56 38	98.33 15	89.60 43	96.79 32	97.65 22	81.83 78	91.52 39	97.23 24	87.94 18	98.91 42	71.31 85	98.37 17	98.17 34
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
IterMVS-LS		92.29 48	91.94 43	93.34 66	96.25 53	86.97 62	96.57 39	97.05 41	90.67 21	89.50 50	94.80 64	86.59 23	97.64 71	89.91 28	86.11 81	95.40 59

PVSNet	86.66 18	92.24 49	91.74 46	93.73 62	97.77 25	83.69 75	92.88 80	96.72 53	87.91 52	93.00 26	94.86 63	78.51 55	99.05 38	86.53 48	97.45 29	98.47 27

XXY-MVS		92.16 50	91.23 50	94.95 47	94.75 75	90.94 33	97.47 19	97.43 31	89.14 37	88.90 56	96.43 42	79.71 51	98.24 57	89.56 30	87.68 75	95.67 57

PatchmatchNet		91.91 51	91.35 49	93.59 64	95.38 61	84.11 72	93.15 77	95.39 71	89.54 32	92.10 35	93.68 74	82.82 41	98.13 60	84.81 56	95.32 43	98.52 22
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Pablo Speciale, Marc Pollefeys: PatchmatchNet: Learned Multi-View Patchmatch Stereo.
FMVSNet3		91.78 52	90.69 54	95.03 43	96.53 46	92.27 14	97.02 29	96.93 46	89.79 31	89.35 52	94.65 68	77.01 59	97.47 74	86.12 51	88.82 70	95.35 61

PCF-MVS	89.48 11	91.56 53	89.95 61	96.36 23	96.60 44	92.52 11	92.51 82	97.26 37	79.41 82	88.90 56	96.56 39	84.04 36	99.55 14	77.01 79	97.30 31	97.01 49
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
PAPM		91.52 54	90.30 56	95.20 41	95.30 65	89.83 42	93.38 76	96.85 51	86.26 62	88.59 59	95.80 52	84.88 31	98.15 59	75.67 81	95.93 40	97.63 45

TR-MVS		91.48 55	90.59 55	94.16 58	96.40 50	87.33 59	95.67 53	95.34 74	87.68 55	91.46 40	95.52 57	76.77 60	98.35 55	82.85 66	93.61 51	96.79 52

MSDG		91.42 56	90.24 58	94.96 46	97.15 35	88.91 51	93.69 72	96.32 61	85.72 63	86.93 68	96.47 41	80.24 49	98.98 40	80.57 73	95.05 44	96.98 50

GBi-Net		91.35 57	90.27 57	94.59 50	96.51 47	91.18 28	97.50 17	96.93 46	88.82 45	89.35 52	94.51 69	73.87 66	97.29 77	86.12 51	88.82 70	95.31 63

FMVSNet2		91.31 58	90.08 59	94.99 44	96.51 47	92.21 15	97.41 20	96.95 44	88.82 45	88.62 58	94.75 65	73.87 66	97.42 75	85.20 54	88.55 73	95.35 61

cascas		91.20 59	90.08 59	94.58 52	94.97 70	89.16 49	93.65 74	97.59 23	79.90 81	89.40 51	92.92 77	75.36 64	98.36 54	92.14 15	94.75 45	96.23 55

LTVRE_ROB	88.41 13	90.99 60	89.92 62	94.19 57	96.18 55	89.55 45	96.31 44	97.09 40	87.88 53	85.67 71	95.91 49	78.79 54	98.57 48	81.50 71	89.98 65	94.44 71
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
RPSCF		90.75 61	90.86 52	90.42 80	96.84 41	76.29 84	95.61 55	96.34 60	83.89 72	91.38 41	97.87 13	76.45 61	98.78 44	87.16 44	92.23 57	96.20 56

EPMVS		90.70 62	89.81 63	93.37 65	94.73 76	84.21 71	93.67 73	88.02 89	89.50 33	92.38 31	93.49 75	77.82 58	97.78 69	86.03 53	92.68 55	98.11 36

ACMH	87.59 16	90.53 63	89.42 65	93.87 60	96.21 54	87.92 55	97.24 25	96.94 45	88.45 47	83.91 75	96.27 45	71.92 70	98.62 47	84.43 58	89.43 68	95.05 67
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
COLMAP_ROB	87.81 15	90.40 64	89.28 66	93.79 61	97.95 23	87.13 61	96.92 31	95.89 67	82.83 76	86.88 69	97.18 25	73.77 68	99.29 26	78.44 76	93.62 50	94.95 68
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
ACMH+	87.92 14	90.20 65	89.18 67	93.25 67	96.48 49	86.45 65	96.99 30	96.68 55	88.83 44	84.79 72	96.22 46	70.16 74	98.53 50	84.42 59	88.04 74	94.77 70

IterMVS		90.15 66	89.67 64	91.61 72	95.48 59	83.72 74	94.33 67	96.12 63	89.99 26	87.31 66	94.15 71	75.78 63	96.27 82	86.97 45	86.89 80	94.83 69

IB-MVS	87.33 17	89.91 67	88.28 69	94.79 49	95.26 67	87.70 58	95.12 64	93.95 84	89.35 34	87.03 67	92.49 79	70.74 73	99.19 30	89.18 33	81.37 87	97.49 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
FMVSNet1		89.88 68	88.31 68	94.59 50	95.41 60	91.18 28	97.50 17	96.93 46	86.62 60	87.41 65	94.51 69	65.94 79	97.29 77	83.04 64	87.43 77	95.31 63

USDC		88.94 69	87.83 71	92.27 69	94.66 77	84.96 68	93.86 71	95.90 66	87.34 57	83.40 77	95.56 56	67.43 77	98.19 58	82.64 68	89.67 67	93.66 76

dp		88.90 70	88.26 70	90.81 77	94.58 78	76.62 83	92.85 81	94.93 78	85.12 66	90.07 46	93.07 76	75.81 62	98.12 62	80.53 74	87.42 78	97.71 43

MIMVSNet		88.50 71	86.76 73	93.72 63	94.84 73	87.77 57	91.39 84	94.05 81	86.41 61	87.99 63	92.59 78	63.27 83	95.82 84	77.44 77	92.84 54	97.57 47

LF4IMVS		87.94 72	87.25 72	89.98 82	92.38 82	80.05 81	94.38 66	95.25 77	87.59 56	84.34 73	94.74 66	64.31 82	97.66 70	84.83 55	87.45 76	92.23 82

FMVSNet5		87.29 73	85.79 74	91.78 71	94.80 74	87.28 60	95.49 58	95.28 75	84.09 70	83.85 76	91.82 82	62.95 84	94.17 87	78.48 75	85.34 82	93.91 75

TinyColmap		86.82 74	85.35 75	91.21 74	94.91 72	82.99 76	93.94 70	94.02 83	83.58 75	81.56 80	94.68 67	62.34 85	98.13 60	75.78 80	87.35 79	92.52 80

TDRefinement		86.53 75	84.76 78	91.85 70	82.23 90	84.25 70	96.38 41	95.35 72	84.97 67	84.09 74	94.94 61	65.76 80	98.34 56	84.60 57	74.52 89	92.97 79

test_0402		86.46 76	84.79 77	91.45 73	95.02 69	85.55 66	96.29 46	94.89 79	80.90 80	82.21 79	93.97 72	68.21 76	97.29 77	62.98 89	88.68 72	91.51 85

UnsupCasMVSNet_eth		85.99 77	84.45 79	90.62 78	89.97 85	82.40 78	93.62 75	97.37 33	89.86 28	78.59 85	92.37 80	65.25 81	95.35 85	82.27 70	70.75 92	94.10 73

CMPMVS	62.92 21	85.62 78	84.92 76	87.74 84	89.14 88	73.12 87	94.17 68	96.80 52	73.98 87	73.65 89	94.93 62	66.36 78	97.61 73	83.95 61	91.28 61	92.48 81
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
PVSNet_0	82.17 19	85.46 79	83.64 80	90.92 75	95.27 66	79.49 82	90.55 87	95.60 69	83.76 74	83.00 78	89.95 85	71.09 71	97.97 64	82.75 67	60.79 94	95.31 63

MIMVSNet1		84.93 80	83.05 81	90.56 79	89.56 87	84.84 69	95.40 60	95.35 72	83.91 71	80.38 83	92.21 81	57.23 88	93.34 88	70.69 86	82.75 86	93.50 77

OpenMVS_ROB	81.14 20	84.42 81	82.28 82	90.83 76	90.06 84	84.05 73	95.73 52	94.04 82	73.89 88	80.17 84	91.53 84	59.15 87	97.64 71	66.92 88	89.05 69	90.80 86

new_pmnet		82.89 82	81.12 84	88.18 83	89.63 86	80.18 80	91.77 83	92.57 86	76.79 86	75.56 88	88.23 87	61.22 86	94.48 86	71.43 84	82.92 84	89.87 87

MVS-HIRNet		82.47 83	81.21 83	86.26 85	95.38 61	69.21 89	88.96 88	89.49 88	66.28 90	80.79 82	74.08 92	68.48 75	97.39 76	71.93 83	95.47 42	92.18 83

UnsupCasMVSNet_bld		82.13 84	79.46 85	90.14 81	88.00 89	82.47 77	90.89 86	96.62 58	78.94 83	75.61 87	84.40 90	56.63 89	96.31 81	77.30 78	66.77 93	91.63 84

HyFIR lowres test		81.01 85	79.31 86	86.13 86	93.19 79	74.03 85	86.45 89	93.06 85	68.57 89	76.44 86	87.03 89	43.72 91	96.27 82	73.27 82	92.05 58	88.09 88

N_pmnet		78.73 86	78.71 87	78.79 87	92.80 80	46.50 97	94.14 69	43.71 96	78.61 84	80.83 81	91.66 83	74.94 65	96.36 80	67.24 87	84.45 83	93.50 77

FPMVS		71.27 87	69.85 88	75.50 89	74.64 91	59.03 93	91.30 85	91.50 87	58.80 92	57.92 91	88.28 86	29.98 93	85.53 91	53.43 92	82.84 85	81.95 90

Gipuma		67.86 88	65.41 89	75.18 90	92.66 81	73.45 86	66.50 95	94.52 80	53.33 93	57.80 92	66.07 95	30.81 92	89.20 89	48.15 95	78.88 88	62.90 96
S. Galliani, K. Lasinger, K. Schindler: Massively Parallel Multiview Stereopsis by Surface Normal Diffusion. ICCV 2015
ANet_high		63.94 89	59.58 90	77.02 88	61.24 96	66.06 90	85.66 90	87.93 90	78.53 85	42.94 93	71.04 93	25.42 94	80.71 93	52.60 93	30.83 98	84.28 89

PNet_i23d		59.01 90	55.87 91	68.44 93	73.98 92	51.37 95	81.36 91	82.41 93	52.37 94	42.49 94	70.39 94	11.39 95	79.99 95	49.77 94	38.71 96	73.97 93

PMVS	53.92 22	58.58 91	55.40 92	68.12 94	51.00 97	48.64 96	78.86 93	87.10 91	46.77 96	35.84 96	74.28 91	8.76 96	86.34 90	42.07 96	73.91 90	69.38 94
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
wuykxyi23d		56.92 92	51.11 93	74.38 92	62.30 95	61.47 92	80.09 92	84.87 92	49.62 95	30.80 97	57.20 97	7.03 97	82.94 92	55.69 91	32.36 97	78.72 92

MVE	50.73 23	53.25 93	48.81 94	66.58 95	65.34 94	57.50 94	72.49 94	70.94 95	40.15 97	39.28 95	63.51 96	6.89 99	73.48 96	38.29 97	42.38 95	68.76 95
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
wuyk23d		25.11 94	24.57 95	26.74 96	73.98 92	39.89 98	57.88 96	9.80 97	12.27 98	10.39 98	6.97 98	7.03 97	36.44 97	25.43 98	17.39 99	3.89 97

ab-mvs-re		8.06 95	10.74 96	0.00 97	0.00 98	0.00 99	0.00 97	0.00 98	0.00 99	0.00 99	96.69 34	0.00 100	0.00 98	0.00 99	0.00 100	0.00 98

ACMMP++															91.02 62

ACMMP++_ref															90.30 64

DeepMVS_CX				74.68 91	90.84 83	64.34 91		81.61 94	65.34 91	67.47 90	88.01 88	48.60 90	80.13 94	62.33 90	73.68 91	79.58 91

Test By Simon												88.73 14

MDTV_nov1_ep13			90.76 53		95.22 68	80.33 79	93.03 79	95.28 75	88.14 49	92.84 30	93.83 73	81.34 45	98.08 63	82.86 65	94.34 48

ITE_SJBPF				92.43 68	95.34 63	85.37 67		95.92 65	91.47 18	87.75 64	96.39 43	71.00 72	97.96 65	82.36 69	89.86 66	93.97 74

MDTV_nov1_ep13_2view						70.35 88	93.10 78		83.88 73	93.55 21		82.47 42		86.25 49		98.38 31

High-resolution multi-view benchmark