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.

DELS-MVS		90.05 1	90.09 2	89.94 1	93.14 19	73.88 2	97.01 1	94.40 13	88.32 1	85.71 7	94.91 13	74.11 3	98.91 2	87.26 6	95.94 1	97.03 2

DeepPCF-MVS	81.17 1	89.72 2	91.38 1	84.72 44	93.00 21	58.16 100	96.72 2	94.41 12	86.50 3	90.25 1	97.83 1	75.46 1	98.67 7	92.78 1	95.49 3	97.32 1

NCCC		89.07 3	89.46 3	87.91 5	96.60 3	69.05 22	96.38 4	94.64 7	84.42 5	86.74 5	96.20 3	66.56 17	98.76 6	89.03 3	94.56 5	95.92 12

DeepC-MVS_fast	79.48 2	87.95 4	88.00 5	87.79 7	95.86 9	68.32 28	95.74 7	94.11 16	83.82 6	83.49 12	96.19 4	64.53 24	98.44 10	83.42 17	94.88 4	96.61 4
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
EPNet		87.84 5	88.38 4	86.23 30	93.30 17	66.05 47	95.26 12	94.84 6	87.09 2	88.06 2	94.53 15	66.79 16	97.34 22	83.89 15	91.68 17	95.29 19

lupinMVS		87.74 6	87.77 6	87.63 9	89.24 57	71.18 8	96.57 3	92.90 29	82.70 8	87.13 3	95.27 8	64.99 20	95.80 42	89.34 2	91.80 15	95.93 11

MG-MVS		87.11 7	86.27 10	89.62 2	97.79 1	76.27 1	94.96 16	94.49 9	78.74 25	83.87 11	92.94 31	64.34 25	96.94 29	75.19 34	94.09 7	95.66 13

CSCG		86.87 8	86.26 11	88.72 3	95.05 12	70.79 10	93.83 23	95.33 1	68.48 82	77.63 28	94.35 16	73.04 4	98.45 9	84.92 13	93.71 9	96.92 3

PHI-MVS		86.83 9	86.85 9	86.78 20	93.47 16	65.55 53	95.39 11	95.10 4	71.77 62	85.69 8	96.52 2	62.07 36	98.77 5	86.06 10	95.60 2	96.03 8

SteuartSystems-ACMMP		86.82 10	86.90 7	86.58 23	90.42 43	66.38 45	96.09 6	93.87 17	77.73 29	84.01 10	95.66 6	63.39 30	97.94 15	87.40 5	93.55 11	95.42 15
Skip Steuart: Steuart Systems R&D Blog.
PVSNet_Blended		86.73 11	86.86 8	86.31 29	93.76 14	67.53 38	96.33 5	93.61 22	82.34 9	81.00 15	93.08 28	63.19 31	97.29 23	87.08 7	91.38 19	94.13 33

jason		86.40 12	86.17 12	87.11 18	86.16 84	70.54 12	95.71 8	92.19 37	82.00 14	84.58 9	94.34 17	61.86 37	95.53 50	87.76 4	90.89 22	95.27 20
jason: jason.
WTY-MVS		86.32 13	85.81 14	87.85 6	92.82 25	69.37 20	95.20 13	95.25 2	82.71 7	81.91 14	94.73 14	67.93 12	97.63 18	79.55 23	82.25 49	96.54 5

MVS_111021_HR		86.19 14	85.80 15	87.37 13	93.17 18	69.79 18	93.99 22	93.76 18	79.08 22	78.88 23	93.99 19	62.25 35	98.15 13	85.93 11	91.15 21	94.15 32

APD-MVS		85.93 15	85.99 13	85.76 37	95.98 8	65.21 56	93.59 28	92.58 33	66.54 85	86.17 6	95.88 5	63.83 28	97.00 26	86.39 9	92.94 13	95.06 23

PAPM		85.89 16	85.46 16	87.18 15	88.20 75	72.42 3	92.41 40	92.77 31	82.11 12	80.34 17	93.07 29	68.27 11	95.02 53	78.39 27	93.59 10	94.09 36

DeepC-MVS	77.85 3	85.52 17	85.24 17	86.37 26	88.80 65	66.64 43	92.15 43	93.68 20	81.07 15	76.91 34	93.64 22	62.59 34	98.44 10	85.50 12	92.84 14	94.03 38
Andreas Kuhn, Christian Sormann, Mattia Rossi, Oliver Erdler, Friedrich Fraundorfer: DeepC-MVS: Deep Confidence Prediction for Multi-View Stereo Reconstruction. 3DV 2020
PAPR		85.15 18	84.47 18	87.18 15	96.02 7	68.29 29	91.85 54	93.00 28	76.59 35	79.03 22	95.00 9	61.59 38	97.61 19	78.16 28	89.00 29	95.63 14

HY-MVS	76.49 5	84.28 19	83.36 23	87.02 19	92.22 29	67.74 35	84.65 90	94.50 8	79.15 20	82.23 13	87.93 68	66.88 15	96.94 29	80.53 21	82.20 50	96.39 6

MAR-MVS		84.18 20	83.43 20	86.44 24	96.25 5	65.93 49	94.28 18	94.27 15	74.41 42	79.16 21	95.61 7	53.99 64	98.88 4	69.62 52	93.26 12	94.50 29
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
PVSNet_Blended_VisFu		83.97 21	83.50 19	85.39 39	90.02 46	66.59 44	93.77 24	91.73 44	77.43 32	77.08 33	89.81 60	63.77 29	96.97 27	79.67 22	88.21 31	92.60 54

PatchFormer_test		83.95 22	82.80 26	87.41 11	92.90 23	70.07 15	89.12 75	94.42 10	82.15 10	77.64 26	91.77 42	70.81 7	96.22 35	65.03 67	81.36 52	95.94 9

MVSFormer		83.75 23	82.88 25	86.37 26	89.24 57	71.18 8	89.07 77	90.69 59	65.80 88	87.13 3	94.34 17	64.99 20	92.67 90	72.83 38	91.80 15	95.27 20

CP-MVS		83.71 24	83.40 22	84.65 45	93.14 19	63.84 67	94.59 17	92.28 35	71.03 66	77.41 30	94.92 12	55.21 57	96.19 37	81.32 20	90.70 23	93.91 39

PVSNet_BlendedMVS		83.38 25	83.43 20	83.22 62	93.76 14	67.53 38	94.06 19	93.61 22	79.13 21	81.00 15	85.14 79	63.19 31	97.29 23	87.08 7	73.91 82	84.83 104

HPM-MVS		83.25 26	82.95 24	84.17 49	92.25 28	62.88 78	90.91 59	91.86 42	70.30 72	77.12 32	93.96 20	56.75 49	96.28 34	82.04 18	91.34 20	93.34 44

PatchFormer-LS_test		83.14 27	81.81 32	87.12 17	92.34 27	69.92 17	88.64 80	93.32 25	82.07 13	74.87 41	91.62 47	68.91 10	96.08 40	66.07 62	78.45 64	95.37 16

PAPM_NR		82.97 28	81.84 31	86.37 26	94.10 13	66.76 42	87.66 86	92.84 30	69.96 75	74.07 47	93.57 23	63.10 33	97.50 20	70.66 47	90.58 25	94.85 24

CLD-MVS		82.73 29	82.35 28	83.86 52	87.90 78	67.65 37	95.45 10	92.18 38	85.06 4	72.58 55	92.27 38	52.46 66	95.78 43	84.18 14	79.06 62	88.16 84
Zhaoxin Li, Wangmeng Zuo, Zhaoqi Wang, Lei Zhang: Confidence-based Large-scale Dense Multi-view Stereo. IEEE Transaction on Image Processing, 2020
Test4		82.73 29	81.18 36	87.41 11	92.90 23	70.07 15	83.48 96	94.42 10	82.15 10	77.64 26	82.78 85	70.81 7	96.22 35	65.03 67	81.36 52	95.94 9

DP-MVS Recon		82.73 29	81.65 33	85.98 32	97.31 2	67.06 40	95.15 15	91.99 39	69.08 80	76.50 36	93.89 21	54.48 60	98.20 12	70.76 46	85.66 40	92.69 51

sss		82.71 32	82.38 27	83.73 56	89.25 55	59.58 96	92.24 42	94.89 5	77.96 27	79.86 20	92.38 36	56.70 50	97.05 25	77.26 30	80.86 56	94.55 28

3Dnovator	73.91 6	82.69 33	80.82 38	88.31 4	89.57 51	71.26 7	92.60 38	94.39 14	78.84 24	67.89 82	92.48 35	48.42 77	98.52 8	68.80 55	94.40 6	95.15 22

MVSTER		82.47 34	82.05 29	83.74 54	92.68 26	69.01 23	91.90 52	93.21 26	79.83 16	72.14 60	85.71 77	74.72 2	94.72 61	75.72 32	72.49 89	87.50 86

CostFormer		82.33 35	81.15 37	85.86 34	89.01 61	68.46 26	82.39 101	93.01 27	75.59 38	80.25 18	81.57 91	72.03 6	94.96 55	79.06 24	77.48 68	94.16 31

API-MVS		82.28 36	80.53 42	87.54 10	96.13 6	70.59 11	93.63 26	91.04 56	65.72 90	75.45 39	92.83 32	56.11 54	98.89 3	64.10 72	89.75 28	93.15 47

IB-MVS	77.80 4	82.18 37	80.46 43	87.35 14	89.14 59	70.28 14	95.59 9	95.17 3	78.85 23	70.19 68	85.82 75	70.66 9	97.67 16	72.19 43	66.52 111	94.09 36
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
3Dnovator+	73.60 7	82.10 38	80.60 41	86.60 22	90.89 40	66.80 41	95.20 13	93.44 24	74.05 45	67.42 84	92.49 34	49.46 74	97.65 17	70.80 45	91.68 17	95.33 17

MVS_111021_LR		82.02 39	81.52 34	83.51 59	88.42 72	62.88 78	89.77 71	88.93 83	76.78 34	75.55 38	93.10 27	50.31 72	95.38 52	83.82 16	87.02 35	92.26 61

PMMVS		81.98 40	82.04 30	81.78 78	89.76 49	56.17 109	91.13 58	90.69 59	77.96 27	80.09 19	93.57 23	46.33 86	94.99 54	81.41 19	87.46 33	94.17 30

EPP-MVSNet		81.79 41	81.52 34	82.61 70	88.77 66	60.21 90	93.02 32	93.66 21	68.52 81	72.90 52	90.39 56	72.19 5	94.96 55	74.93 35	79.29 61	92.67 52

ACMMP		81.49 42	80.67 40	83.93 51	91.71 31	62.90 77	92.13 44	92.22 36	71.79 61	71.68 63	93.49 25	50.32 71	96.96 28	78.47 26	84.22 43	91.93 63
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
CDS-MVSNet		81.43 43	80.74 39	83.52 58	86.26 83	64.45 62	92.09 47	90.65 62	75.83 37	73.95 48	89.81 60	63.97 27	92.91 86	71.27 44	82.82 47	93.20 46
Khang Truong Giang, Soohwan Song, Sungho Jo: Curvature-guided dynamic scale networks for Multi-view Stereo. ICLR 2022
Vis-MVSNet		80.92 44	79.98 44	83.74 54	88.48 69	61.80 82	93.44 30	88.26 94	73.96 46	77.73 25	91.76 44	49.94 73	94.76 59	65.84 65	90.37 26	94.65 27
Jingyang Zhang, Yao Yao, Shiwei Li, Zixin Luo, Tian Fang: Visibility-aware Multiview Stereo Network. BMVC 2020
tpmrst		80.57 45	79.14 53	84.84 42	90.10 45	68.28 30	81.70 103	89.72 74	77.63 31	75.96 37	79.54 104	64.94 22	92.71 89	75.43 33	77.28 71	93.55 42

1112_ss		80.56 46	79.83 45	82.77 67	88.65 67	60.78 86	92.29 41	88.36 91	72.58 54	72.46 57	94.95 10	65.09 19	93.42 80	66.38 61	77.71 66	94.10 35

BH-w/o		80.49 47	79.30 50	84.05 50	90.83 41	64.36 66	93.60 27	89.42 77	74.35 43	69.09 76	90.15 58	55.23 56	95.61 48	64.61 69	86.43 38	92.17 62

TAMVS		80.37 48	79.45 48	83.13 64	85.14 90	63.37 72	91.23 57	90.76 58	74.81 41	72.65 54	88.49 64	60.63 42	92.95 84	69.41 53	81.95 51	93.08 48

HPM-MVS_fast		80.25 49	79.55 47	82.33 73	91.55 33	59.95 92	91.32 56	89.16 79	65.23 92	74.71 42	93.07 29	47.81 83	95.74 44	74.87 36	88.23 30	91.31 67

ab-mvs		80.18 50	78.31 57	85.80 35	88.44 71	65.49 55	83.00 100	92.67 32	71.82 60	77.36 31	85.01 80	54.50 59	96.59 31	76.35 31	75.63 76	95.32 18

IS-MVSNet		80.14 51	79.41 49	82.33 73	87.91 77	60.08 91	91.97 50	88.27 93	72.90 53	71.44 64	91.73 46	61.44 39	93.66 79	62.47 75	86.53 36	93.24 45

PVSNet	73.49 8	80.05 52	78.63 55	84.31 47	90.92 38	64.97 60	92.47 39	91.05 55	79.18 19	72.43 58	90.51 55	37.05 106	94.06 70	68.06 56	86.00 39	93.90 40

QAPM		79.95 53	77.39 63	87.64 8	89.63 50	71.41 6	93.30 31	93.70 19	65.34 91	67.39 85	91.75 45	47.83 82	98.96 1	57.71 87	89.81 27	92.54 56

UGNet		79.87 54	78.68 54	83.45 61	89.96 47	61.51 84	92.13 44	90.79 57	76.83 33	78.85 24	86.33 72	38.16 100	96.17 38	67.93 57	87.17 34	92.67 52
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
tpm2		79.80 55	77.95 58	85.34 40	88.28 73	68.26 31	81.56 105	91.42 48	70.11 73	77.59 29	80.50 96	67.40 13	94.26 67	67.34 59	77.35 69	93.51 43

CPTT-MVS		79.59 56	79.16 52	80.89 85	91.54 34	59.80 94	92.10 46	88.54 90	60.42 109	72.96 51	93.28 26	48.27 78	92.80 87	78.89 25	86.50 37	90.06 76

Test_1112_low_res		79.56 57	78.60 56	82.43 71	88.24 74	60.39 88	92.09 47	87.99 97	72.10 58	71.84 61	87.42 70	64.62 23	93.04 82	65.80 66	77.30 70	93.85 41

BH-RMVSNet		79.46 58	77.65 61	84.89 41	91.68 32	65.66 52	93.55 29	88.09 95	72.93 52	73.37 50	91.12 50	46.20 87	96.12 39	56.28 91	85.61 41	92.91 50

PCF-MVS	73.15 9	79.29 59	77.63 62	84.29 48	86.06 85	65.96 48	87.03 88	91.10 54	69.86 76	69.79 73	90.64 51	57.54 46	96.59 31	64.37 71	82.29 48	90.32 74
Andreas Kuhn, Shan Lin, Oliver Erdler: Plane Completion and Filtering for Multi-View Stereo Reconstruction. GCPR 2019
Vis-MVSNet (Re-imp)		79.24 60	79.57 46	78.24 101	88.46 70	52.29 119	90.41 65	89.12 80	74.24 44	69.13 75	91.91 41	65.77 18	90.09 113	59.00 84	88.09 32	92.33 59

114514_t		79.17 61	77.67 60	83.68 57	95.32 11	65.53 54	92.85 34	91.60 47	63.49 96	67.92 81	90.63 53	46.65 85	95.72 46	67.01 60	83.54 44	89.79 77

AdaColmap		78.94 62	77.00 67	84.76 43	96.34 4	65.86 51	92.66 37	87.97 98	62.18 101	70.56 66	92.37 37	43.53 92	97.35 21	64.50 70	82.86 46	91.05 69

tpmp4_e23		78.85 63	76.55 68	85.77 36	89.25 55	68.39 27	81.63 104	91.38 50	70.40 70	75.21 40	79.22 106	67.37 14	94.79 58	58.98 85	75.51 77	94.13 33

EPNet_dtu		78.80 64	79.26 51	77.43 105	88.06 76	49.71 126	91.96 51	91.95 41	77.67 30	76.56 35	91.28 49	58.51 43	90.20 111	56.37 90	80.95 55	92.39 58

TR-MVS		78.77 65	77.37 64	82.95 65	90.49 42	60.88 85	93.67 25	90.07 68	70.08 74	74.51 43	91.37 48	45.69 88	95.70 47	60.12 77	80.32 57	92.29 60

BH-untuned		78.68 66	77.08 65	83.48 60	89.84 48	63.74 68	92.70 36	88.59 89	71.57 63	66.83 89	88.65 63	51.75 69	95.39 51	59.03 83	84.77 42	91.32 66

OMC-MVS		78.67 67	77.91 59	80.95 84	85.76 88	57.40 105	88.49 81	88.67 87	73.85 47	72.43 58	92.10 39	49.29 75	94.55 63	72.73 40	77.89 65	90.91 70

tpm		78.58 68	77.03 66	83.22 62	85.94 86	64.56 61	83.21 99	91.14 53	78.31 26	73.67 49	79.68 102	64.01 26	92.09 98	66.07 62	71.26 94	93.03 49

OpenMVS	70.45 11	78.54 69	75.92 72	86.41 25	85.93 87	71.68 4	92.74 35	92.51 34	66.49 86	64.56 96	91.96 40	43.88 91	98.10 14	54.61 93	90.65 24	89.44 80

EPMVS		78.49 70	75.98 71	86.02 31	91.21 36	69.68 19	80.23 111	91.20 51	75.25 40	72.48 56	78.11 108	54.65 58	93.69 78	57.66 88	83.04 45	94.69 26

cascas		78.18 71	75.77 73	85.41 38	87.14 81	69.11 21	92.96 33	91.15 52	66.71 84	70.47 67	86.07 73	37.49 103	96.48 33	70.15 49	79.80 59	90.65 71

XXY-MVS		77.94 72	76.44 69	82.43 71	82.60 107	64.44 63	92.01 49	91.83 43	73.59 49	70.00 70	85.82 75	54.43 61	94.76 59	69.63 51	68.02 108	88.10 85

HyFIR		77.94 72	75.03 75	86.66 21	90.91 39	71.56 5	82.25 102	88.91 84	66.06 87	68.16 80	69.11 129	56.65 51	92.27 95	59.29 82	93.97 8	96.14 7

FMVSNet3		77.73 74	76.04 70	82.80 66	91.20 37	68.99 24	91.87 53	91.99 39	73.35 50	67.04 86	83.19 84	56.62 52	92.14 97	59.80 79	69.34 97	87.28 87

PatchmatchNet		77.46 75	74.63 76	85.96 33	89.55 52	70.35 13	79.97 113	89.55 76	72.23 57	70.94 65	76.91 114	57.03 48	92.79 88	54.27 96	81.17 54	94.74 25
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Pablo Speciale, Marc Pollefeys: PatchmatchNet: Learned Multi-View Patchmatch Stereo.
IterMVS-LS		76.49 76	75.18 74	80.43 86	84.49 95	62.74 80	90.64 62	88.80 85	72.40 56	65.16 92	81.72 90	60.98 40	92.27 95	67.74 58	64.65 114	86.29 94
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Marc Pollefeys: IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo.
FMVSNet2		76.07 77	74.01 79	82.26 76	88.85 62	67.66 36	91.33 55	91.61 46	70.84 67	65.98 90	82.25 87	48.03 79	92.00 99	58.46 86	68.73 103	87.10 88

PT_06_test		75.88 78	73.82 82	82.06 77	83.64 103	65.93 49	89.12 75	91.40 49	75.33 39	64.50 97	91.77 42	51.90 67	94.00 73	54.55 94	62.20 117	86.26 95

LPG-MVS_test		75.82 79	74.58 77	79.56 95	84.31 98	59.37 97	90.44 64	89.73 72	69.49 78	64.86 93	88.42 65	38.65 97	94.30 65	72.56 41	72.76 86	85.01 102

test1		75.65 80	73.83 80	81.10 80	88.85 62	65.11 57	90.01 67	90.32 63	70.84 67	67.04 86	80.25 99	48.03 79	91.54 103	59.80 79	69.34 97	86.64 90

GBi-Net		75.65 80	73.83 80	81.10 80	88.85 62	65.11 57	90.01 67	90.32 63	70.84 67	67.04 86	80.25 99	48.03 79	91.54 103	59.80 79	69.34 97	86.64 90

ACMP	71.68 10	75.58 82	74.23 78	79.62 93	84.97 92	59.64 95	90.80 61	89.07 81	70.39 71	62.95 104	87.30 71	38.28 99	93.87 76	72.89 37	71.45 92	85.36 101
Qingshan Xu and Wenbing Tao: Planar Prior Assisted PatchMatch Multi-View Stereo. AAAI 2020
tpm cat1		75.30 83	72.21 91	84.58 46	88.52 68	67.77 34	78.16 120	88.02 96	61.88 103	68.45 79	76.37 115	60.65 41	94.03 72	53.77 99	74.11 80	91.93 63

PLC	68.80 14	75.23 84	73.68 83	79.86 91	92.93 22	58.68 99	90.64 62	88.30 92	60.90 106	64.43 98	90.53 54	42.38 94	94.57 62	56.52 89	76.54 72	86.33 93
Jie Liao, Yanping Fu, Qingan Yan, Chunxia xiao: Pyramid Multi-View Stereo with Local Consistency. Pacific Graphics 2019
dp		75.01 85	72.09 92	83.76 53	89.28 54	66.22 46	79.96 114	89.75 71	71.16 65	67.80 83	77.19 113	51.81 68	92.54 92	50.39 106	71.44 93	92.51 57

TAPA-MVS	70.22 12	74.94 86	73.53 84	79.17 97	90.40 44	52.07 120	89.19 74	89.61 75	62.69 100	70.07 69	92.67 33	48.89 76	94.32 64	38.26 126	79.97 58	91.12 68
Andrea Romanoni, Matteo Matteucci: TAPA-MVS: Textureless-Aware PAtchMatch Multi-View Stereo. ICCV 2019
XVG-OURS-SEG-HR		74.70 87	73.08 85	79.57 94	78.25 120	57.33 106	80.49 108	87.32 101	63.22 97	68.76 77	90.12 59	44.89 90	91.59 102	70.55 48	74.09 81	89.79 77

ACMM	69.62 13	74.34 88	72.73 86	79.17 97	84.25 100	57.87 102	90.36 66	89.93 69	63.17 98	65.64 91	86.04 74	37.79 102	94.10 69	65.89 64	71.52 91	85.55 100
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
CNLPA		74.31 89	72.30 90	80.32 87	91.49 35	61.66 83	90.85 60	80.72 120	56.67 116	63.85 101	90.64 51	46.75 84	90.84 107	53.79 98	75.99 75	88.47 83

XVG-OURS		74.25 90	72.46 89	79.63 92	78.45 119	57.59 104	80.33 109	87.39 100	63.86 95	68.76 77	89.62 62	40.50 95	91.72 100	69.00 54	74.25 79	89.58 79

CR-MVSNet		73.79 91	70.82 94	82.70 68	83.15 105	67.96 32	70.25 126	84.00 112	73.67 48	69.97 71	72.41 124	57.82 44	89.48 116	52.99 103	73.13 84	90.64 72

test_djsdf		73.76 92	72.56 88	77.39 106	77.00 123	53.93 115	89.07 77	90.69 59	65.80 88	63.92 99	82.03 89	43.14 93	92.67 90	72.83 38	68.53 104	85.57 99

FMVSNet1		72.71 93	69.91 96	81.10 80	83.60 104	65.11 57	90.01 67	90.32 63	63.92 94	63.56 103	80.25 99	36.35 107	91.54 103	54.46 95	66.75 109	86.64 90

IterMVS		72.65 94	70.83 93	78.09 102	82.17 108	62.96 74	87.64 87	86.28 103	71.56 64	60.44 105	78.85 107	45.42 89	86.66 126	63.30 73	61.83 118	84.65 105
Fangjinhua Wang, Silvano Galliani, Christoph Vogel, Marc Pollefeys: IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo.
PatchMatch-RL		72.06 95	69.98 95	78.28 100	89.51 53	55.70 112	83.49 95	83.39 116	61.24 105	63.72 102	82.76 86	34.77 110	93.03 83	53.37 102	77.59 67	86.12 96

PVSNet_0	68.08 15	71.81 96	68.32 99	82.27 75	84.68 93	62.31 81	88.68 79	90.31 66	75.84 36	57.93 109	80.65 95	37.85 101	94.19 68	69.94 50	29.05 143	90.31 75

MIMVSNet		71.64 97	68.44 97	81.23 79	81.97 111	64.44 63	73.05 124	88.80 85	69.67 77	64.59 95	74.79 120	32.79 113	87.82 123	53.99 97	76.35 73	91.42 65

F-COLMAP		70.66 98	68.44 97	77.32 107	86.37 82	55.91 111	88.00 83	86.32 102	56.94 115	57.28 111	88.07 67	33.58 112	92.49 93	51.02 105	68.37 105	83.55 107

DP-MVS		69.90 99	66.48 100	80.14 89	95.36 10	62.93 75	89.56 72	76.11 126	50.27 124	57.69 110	85.23 78	39.68 96	95.73 45	33.35 131	71.05 95	81.78 116

RPMNet		69.58 100	65.21 107	82.70 68	83.15 105	67.96 32	70.25 126	86.15 105	46.83 132	69.97 71	65.10 133	56.48 53	89.48 116	35.79 129	73.13 84	90.64 72

MSDG		69.54 101	65.73 102	80.96 83	85.11 91	63.71 69	84.19 92	83.28 117	56.95 114	54.50 114	84.03 82	31.50 118	96.03 41	42.87 120	69.13 100	83.14 111

LS3D		69.17 102	66.40 101	77.50 104	91.92 30	56.12 110	85.12 89	80.37 121	46.96 130	56.50 112	87.51 69	37.25 104	93.71 77	32.52 136	79.40 60	82.68 113

PatchT		69.11 103	65.37 106	80.32 87	82.07 110	63.68 70	67.96 131	87.62 99	50.86 123	69.37 74	65.18 132	57.09 47	88.53 121	41.59 122	66.60 110	88.74 81

ACMH	63.93 17	68.62 104	64.81 108	80.03 90	85.22 89	63.25 73	87.72 85	84.66 110	60.83 107	51.57 121	79.43 105	27.29 126	94.96 55	41.76 121	64.84 113	81.88 115
Qingshan Xu and Wenbing Tao: Multi-Scale Geometric Consistency Guided Multi-View Stereo. CVPR 2019
EG-PatchMatch MVS		68.55 105	65.41 105	77.96 103	78.69 118	62.93 75	89.86 70	89.17 78	60.55 108	50.27 123	77.73 110	22.60 133	94.06 70	47.18 114	72.65 88	76.88 128

XVG-ACMP-BASELINE		68.04 106	65.53 104	75.56 112	74.06 130	52.37 118	78.43 117	85.88 107	62.03 102	58.91 108	81.21 93	20.38 135	91.15 106	60.69 76	68.18 106	83.16 110

FMVSNet5		68.04 106	65.66 103	75.18 113	84.43 96	57.89 101	83.54 94	86.26 104	61.83 104	53.64 117	73.30 123	37.15 105	85.08 127	48.99 110	61.77 119	82.56 114

ACMH+	65.35 16	67.65 108	64.55 109	76.96 108	84.59 94	57.10 107	88.08 82	80.79 119	58.59 112	53.00 118	81.09 94	26.63 128	92.95 84	46.51 115	61.69 120	80.82 119

Patchmtry		67.53 109	63.93 112	78.34 99	82.12 109	64.38 65	68.72 128	84.00 112	48.23 129	59.24 106	72.41 124	57.82 44	89.27 118	46.10 116	56.68 125	81.36 117

USDC		67.43 110	64.51 110	76.19 111	77.94 121	55.29 113	78.38 118	85.00 109	73.17 51	48.36 127	80.37 97	21.23 134	92.48 94	52.15 104	64.02 115	80.81 120

CMPMVS	48.56 21	66.77 111	64.41 111	73.84 119	70.65 134	50.31 123	77.79 121	85.73 108	45.54 133	44.76 134	82.14 88	35.40 109	90.14 112	63.18 74	74.54 78	81.07 118
M. Jancosek, T. Pajdla: Multi-View Reconstruction Preserving Weakly-Supported Surfaces. CVPR 2011
OpenMVS_ROB	61.12 18	66.39 112	62.92 115	76.80 110	76.51 124	57.77 103	89.22 73	83.41 115	55.48 118	53.86 116	77.84 109	26.28 129	93.95 74	34.90 130	68.76 102	78.68 125

LTVRE_ROB	59.60 19	66.27 113	63.54 113	74.45 115	84.00 102	51.55 122	67.08 132	83.53 114	58.78 111	54.94 113	80.31 98	34.54 111	93.23 81	40.64 124	68.03 107	78.58 126
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
JIA-IIPM		66.06 114	62.45 116	76.88 109	81.42 113	54.45 114	57.49 139	88.67 87	49.36 125	63.86 100	46.86 140	56.06 55	90.25 110	49.53 109	68.83 101	85.95 98

UnsupCasMVSNet_eth		65.79 115	63.10 114	73.88 118	70.71 133	50.29 124	81.09 106	89.88 70	72.58 54	49.25 125	74.77 121	32.57 114	87.43 125	55.96 92	41.04 138	83.90 106

SixPastTwentyTwo		64.92 116	61.78 119	74.34 117	78.74 117	49.76 125	83.42 97	79.51 123	62.86 99	50.27 123	77.35 111	30.92 121	90.49 108	45.89 117	47.06 133	82.78 112

OurMVSNet-221017-0		64.68 117	62.17 117	72.21 124	76.08 127	47.35 130	80.67 107	81.02 118	56.19 117	51.60 120	79.66 103	27.05 127	88.56 120	53.60 100	53.63 130	80.71 121

test_0402		64.54 118	61.09 120	74.92 114	84.10 101	60.75 87	87.95 84	79.71 122	52.03 122	52.41 119	77.20 112	32.21 116	91.64 101	23.14 139	61.03 122	72.36 132

RPSCF		64.24 119	61.98 118	71.01 127	76.10 126	45.00 133	75.83 123	75.94 127	46.94 131	58.96 107	84.59 81	31.40 119	82.00 132	47.76 112	60.33 124	86.04 97

K. test v3		63.09 120	59.61 122	73.53 121	76.26 125	49.38 127	83.27 98	77.15 125	64.35 93	47.77 128	72.32 126	28.73 123	87.79 124	49.93 108	36.69 141	83.41 109

COLMAP_ROB	57.96 20	62.98 121	59.65 121	72.98 123	81.44 112	53.00 117	83.75 93	75.53 130	48.34 128	48.81 126	81.40 92	24.14 130	90.30 109	32.95 133	60.52 123	75.65 130
Johannes L. Schönberger, Enliang Zheng, Marc Pollefeys, Jan-Michael Frahm: Pixelwise View Selection for Unstructured Multi-View Stereo. ECCV 2016
UnsupCasMVSNet_bld		61.60 122	57.71 123	73.29 122	68.73 136	51.64 121	78.61 116	89.05 82	57.20 113	46.11 130	61.96 135	28.70 124	88.60 119	50.08 107	38.90 139	79.63 124

TinyColmap		60.32 123	56.42 126	72.00 125	78.78 116	53.18 116	78.36 119	75.64 128	52.30 121	41.59 139	75.82 118	14.76 139	88.35 122	35.84 128	54.71 129	74.46 131

MVS-HIRNet		60.25 124	55.55 127	74.35 116	84.37 97	56.57 108	71.64 125	74.11 132	34.44 141	45.54 133	42.24 141	31.11 120	89.81 114	40.36 125	76.10 74	76.67 129

MIMVSNet1		60.16 125	57.33 124	68.67 129	69.71 135	44.13 134	78.92 115	84.21 111	55.05 119	44.63 135	71.85 127	23.91 131	81.54 133	32.63 135	55.03 128	80.35 122

testpf		57.17 126	56.93 125	57.88 136	79.13 115	42.40 135	34.23 143	85.97 106	52.64 120	47.66 129	66.50 131	36.33 108	79.65 136	53.60 100	56.31 126	51.60 141

DSMNet-mixed		56.78 127	54.44 128	63.79 132	63.21 137	29.44 142	64.43 135	64.10 140	42.12 139	51.32 122	71.60 128	31.76 117	75.04 137	36.23 127	65.20 112	86.87 89

LP		56.71 128	51.64 130	71.91 126	80.08 114	60.33 89	61.72 137	75.61 129	43.87 137	43.76 137	60.30 137	30.46 122	84.05 128	22.94 140	46.06 135	71.34 133

TDRefinement		55.28 129	51.58 131	66.39 130	59.53 140	46.15 132	76.23 122	72.80 133	44.60 136	42.49 138	76.28 116	15.29 137	82.39 131	33.20 132	43.75 136	70.62 135

HyFIR lowres test		54.68 130	50.80 132	66.34 131	74.15 129	49.02 128	53.99 140	62.10 142	44.68 135	37.46 140	63.99 134	13.84 141	80.23 134	43.01 119	63.99 116	69.76 136

LF4IMVS		54.01 131	52.12 129	59.69 134	62.41 139	39.91 138	68.59 129	68.28 137	42.96 138	44.55 136	75.18 119	14.09 140	68.39 140	41.36 123	51.68 131	70.78 134

N_pmnet		50.55 132	49.11 133	54.88 138	77.17 122	4.02 149	84.36 91	2.00 149	48.59 126	45.86 131	68.82 130	32.22 115	82.80 130	31.58 137	51.38 132	77.81 127

new_pmnet		49.31 133	46.44 134	57.93 135	62.84 138	40.74 137	68.47 130	62.96 141	36.48 140	35.09 141	57.81 138	14.97 138	72.18 138	32.86 134	46.44 134	60.88 139

FPMVS		45.64 134	43.10 135	53.23 139	51.42 142	36.46 140	64.97 134	71.91 135	29.13 143	27.53 142	61.55 136	9.83 143	65.01 142	16.00 145	55.58 127	58.22 140

no-one		44.13 135	38.39 136	61.34 133	45.91 145	41.94 136	61.67 138	75.07 131	45.05 134	20.07 143	40.68 144	11.58 142	79.82 135	30.18 138	15.30 145	62.26 138

ANet_high		40.27 136	35.20 137	55.47 137	34.74 146	34.47 141	63.84 136	71.56 136	48.42 127	18.80 144	41.08 142	9.52 144	64.45 143	20.18 142	8.66 149	67.49 137

Gipuma		34.91 137	31.44 138	45.30 141	70.99 132	39.64 139	19.85 146	72.56 134	20.10 146	16.16 146	21.47 147	5.08 146	71.16 139	13.07 146	43.70 137	25.08 147
S. Galliani, K. Lasinger, K. Schindler: Massively Parallel Multiview Stereopsis by Surface Normal Diffusion. ICCV 2015
PNet_i23d		32.77 138	29.98 139	41.11 143	48.05 143	29.17 143	65.82 133	50.02 145	21.42 144	14.74 147	37.19 145	1.11 149	55.11 145	19.75 143	11.77 147	39.06 143

PMVS	26.43 22	31.84 139	28.16 140	42.89 142	25.87 149	27.58 145	50.92 141	49.78 146	21.37 145	14.17 148	40.81 143	2.01 147	66.62 141	9.61 148	38.88 140	34.49 146
Y. Furukawa, J. Ponce: Accurate, dense, and robust multiview stereopsis. PAMI (2010)
wuykxyi23d		29.03 140	23.09 141	46.84 140	31.67 148	28.82 144	43.46 142	57.72 144	14.39 147	7.52 150	20.84 148	0.64 150	60.29 144	21.57 141	10.04 148	51.40 142

MVE	24.84 23	24.35 141	19.77 142	38.09 144	34.56 147	26.92 146	26.57 144	38.87 147	11.73 148	11.37 149	27.44 146	1.37 148	50.42 146	11.41 147	14.60 146	36.93 144
Simon Fuhrmann, Fabian Langguth, Michael Goesele: MVE - A Multi-View Reconstruction Environment. EUROGRAPHICS Workshops on Graphics and Cultural Heritage (2014)
wuyk23d		11.30 142	10.95 143	12.33 146	48.05 143	19.89 148	25.89 145	1.92 150	3.58 149	3.12 151	1.37 149	0.64 150	15.77 148	6.23 149	7.77 150	1.35 148

ab-mvs-re		7.91 143	10.55 144	0.00 147	0.00 150	0.00 150	0.00 147	0.00 151	0.00 150	0.00 152	94.95 10	0.00 152	0.00 149	0.00 150	0.00 151	0.00 149

DeepMVS_CX				34.71 145	51.45 141	24.73 147		28.48 148	31.46 142	17.49 145	52.75 139	5.80 145	42.60 147	18.18 144	19.42 144	36.81 145

HQP4-MVS										74.18 44			95.61 48			88.63 82

lessismore_v0				73.72 120	72.93 131	47.83 129		61.72 143		45.86 131	73.76 122	28.63 125	89.81 114	47.75 113	31.37 142	83.53 108

LGP-MVS_train				79.56 95	84.31 98	59.37 97		89.73 72	69.49 78	64.86 93	88.42 65	38.65 97	94.30 65	72.56 41	72.76 86	85.01 102

door								66.57 138

HQP5-MVS						63.66 71

HQP-NCC					87.54 79		94.06 19		79.80 17	74.18 44

ACMP_Plane					87.54 79		94.06 19		79.80 17	74.18 44

BP-MVS														77.63 29

HQP3-MVS								91.70 45							78.90 63

ITE_SJBPF				70.43 128	74.44 128	47.06 131		77.32 124	60.16 110	54.04 115	83.53 83	23.30 132	84.01 129	43.07 118	61.58 121	80.21 123

HQP2-MVS												51.63 70

NP-MVS											90.30 57

MDTV_nov1_ep13_2view						59.90 93	80.13 112		67.65 83	72.79 53		54.33 62		59.83 78		92.58 55

MDTV_nov1_ep13			72.61 87		89.06 60	68.48 25	80.33 109	90.11 67	71.84 59	71.81 62	75.92 117	53.01 65	93.92 75	48.04 111	73.38 83

ACMMP++_ref															71.63 90

ACMMP++															69.72 96

HQP-MVS												16.21 136

Test By Simon												54.21 63

door-mid								66.01 139

High-resolution multi-view benchmark