Low-res two-view results

This table lists the benchmark results for the low-res two-view scenario. This benchmark evaluates the Middlebury stereo metrics (for all metrics, smaller is better):

bad 0.5, 1.0, 2.0, 4.0 [%]: Fraction of pixels with errors larger than the given number of disparities.
Average error [px]: The per-pixel average disparity error.
Root mean square error [px]: The per-pixel root mean square disparity error.
50%, 90%, 95%, 99% error quantile [px]: The highest disparity error within the given percentage of best pixels (for 50%, this is the median error).
Time [s]: The runtime of the method.

The mask determines whether the metric is evaluated for all pixels with ground truth, or only for pixels which are visible in both images (non-occluded).
The coverage selector allows to limit the table to results for all pixels (dense), or a given minimum fraction of pixels.

Methods with suffix _ROB may participate in the Robust Vision Challenge.

Click one or more dataset result cells or column headers to show visualizations. Most visualizations are only available for training datasets. The visualizations may not work with mobile browsers.

Method	all	lakes. 1l	lakes. 1s	sand box 1l	sand box 1s	stora. room 1l	stora. room 1s	stora. room 2l	stora. room 2s	stora. room 2 1l	stora. room 2 1s	stora. room 2 2l	stora. room 2 2s	stora. room 3l	stora. room 3s	tunnel 1l	tunnel 1s	tunnel 2l	tunnel 2s	tunnel 3l	tunnel 3s

R-Stereo Train	0.15 1	0.03 1	0.77 4	0.00 1	0.00 1	0.74 36	0.00 1	1.36 54	0.09 2	0.02 31	0.00 1	0.01 1	0.00 1	0.01 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.01 21

R-Stereo	0.15 1	0.03 1	0.77 4	0.00 1	0.00 1	0.74 36	0.00 1	1.36 54	0.09 2	0.02 31	0.00 1	0.01 1	0.00 1	0.01 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.01 21

HITNet	0.19 3	0.55 59	1.36 18	0.00 1	0.00 1	0.25 5	0.00 1	1.09 44	0.23 15	0.00 1	0.00 1	0.14 13	0.02 14	0.02 3	0.14 25	0.00 1	0.00 1	0.00 1	0.00 1	0.05 36	0.00 1

AdaStereo	0.20 4	0.06 3	0.86 6	0.00 1	0.00 1	0.25 5	0.00 1	1.71 68	0.56 27	0.00 1	0.00 1	0.31 27	0.01 8	0.11 7	0.11 21	0.00 1	0.00 1	0.00 1	0.00 1	0.01 10	0.00 1
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
MLCV	0.23 5	0.46 41	1.15 14	0.00 1	0.00 1	0.45 15	0.00 1	0.57 18	0.15 6	0.09 49	0.00 1	0.27 23	0.03 24	1.09 14	0.25 32	0.00 1	0.00 1	0.00 1	0.00 1	0.05 36	0.00 1

DeepPruner_ROB	0.23 5	0.37 33	1.26 15	0.00 1	0.00 1	1.11 65	0.03 58	0.20 4	0.07 1	0.00 1	0.00 1	0.39 36	0.03 24	1.12 15	0.02 7	0.00 1	0.00 1	0.00 1	0.00 1	0.04 32	0.00 1

StereoDRNet-Refined	0.24 7	0.08 8	0.68 2	0.00 1	0.00 1	0.83 43	0.00 1	0.66 22	0.61 46	0.00 1	0.00 1	0.50 43	0.02 14	1.28 17	0.08 18	0.00 1	0.00 1	0.00 1	0.00 1	0.10 57	0.04 56
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
MFMNet_re	0.24 7	0.74 79	2.31 50	0.53 90	0.00 1	0.17 2	0.01 39	0.02 1	0.59 43	0.01 22	0.03 53	0.12 10	0.00 1	0.07 4	0.01 3	0.00 1	0.00 1	0.00 1	0.00 1	0.19 77	0.01 21

iResNet	0.25 9	0.27 19	1.68 30	0.00 1	0.00 1	0.69 33	0.00 1	1.22 51	0.56 27	0.00 1	0.00 1	0.04 4	0.03 24	0.12 8	0.30 35	0.00 1	0.00 1	0.00 1	0.00 1	0.05 36	0.01 21

DN-CSS_ROB	0.26 10	0.75 80	1.33 17	0.00 1	0.00 1	0.99 57	0.00 1	1.04 39	0.61 46	0.01 22	0.00 1	0.19 18	0.02 14	0.07 4	0.01 3	0.00 1	0.00 1	0.00 1	0.00 1	0.22 81	0.01 21

DLCB_ROB	0.27 11	0.16 13	0.71 3	0.00 1	0.00 1	0.44 14	0.00 1	0.43 10	0.57 31	0.02 31	0.04 54	0.63 46	0.04 31	2.18 27	0.24 31	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1

NCCL2	0.29 12	0.50 46	1.52 24	0.00 1	0.00 1	0.81 41	0.00 1	0.45 11	0.22 13	0.00 1	0.00 1	0.02 3	0.03 24	2.14 25	0.06 14	0.00 1	0.00 1	0.02 85	0.00 1	0.01 10	0.03 45

iResNet_ROB	0.30 13	0.66 71	1.41 20	0.00 1	0.00 1	0.70 34	0.00 1	0.79 29	0.57 31	0.04 43	0.01 37	0.08 6	0.01 8	1.55 18	0.03 9	0.00 1	0.00 1	0.00 1	0.00 1	0.06 42	0.00 1

ETE_ROB	0.31 14	0.34 24	0.89 7	0.00 1	0.00 1	0.97 53	0.01 39	1.45 61	0.57 31	0.02 31	0.00 1	0.17 14	0.02 14	1.75 19	0.07 16	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1

DISCO	0.31 14	0.11 9	1.73 33	0.16 80	0.16 81	0.31 7	0.00 1	0.09 2	0.17 8	0.00 1	0.01 37	1.51 71	0.00 1	1.21 16	0.62 48	0.00 1	0.00 1	0.00 1	0.00 1	0.03 28	0.01 21

LALA_ROB	0.31 14	0.38 34	0.91 10	0.00 1	0.01 48	1.00 58	0.01 39	1.60 63	0.61 46	0.15 55	0.00 1	0.09 8	0.01 8	1.04 13	0.41 39	0.00 1	0.00 1	0.00 1	0.00 1	0.02 20	0.00 1

CFNet_RVC	0.31 14	0.34 24	0.96 11	0.00 1	0.00 1	1.08 62	0.00 1	0.19 3	0.15 6	0.01 22	0.00 1	0.48 41	0.00 1	2.95 40	0.05 13	0.00 1	0.00 1	0.01 79	0.00 1	0.01 10	0.01 21

NLCA_NET_v2_RVC	0.35 18	0.49 45	2.00 40	0.00 1	0.02 55	0.96 51	0.00 1	0.55 17	0.18 11	0.03 37	0.00 1	0.35 32	0.10 44	2.37 30	0.01 3	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
iResNetv2_ROB	0.36 19	1.03 88	2.96 63	0.00 1	0.00 1	0.58 22	0.00 1	1.51 62	0.58 39	0.01 22	0.00 1	0.08 6	0.02 14	0.07 4	0.02 7	0.00 1	0.00 1	0.00 1	0.00 1	0.29 89	0.02 35

CC-Net-ROB	0.37 20	0.50 46	2.09 47	0.00 1	0.02 55	0.98 54	0.00 1	0.46 12	0.17 8	0.08 47	0.00 1	0.33 28	0.11 45	2.54 33	0.01 3	0.00 1	0.00 1	0.00 1	0.01 76	0.01 10	0.00 1

XPNet_ROB	0.37 20	0.34 24	1.01 12	0.01 47	0.00 1	0.96 51	0.00 1	1.11 46	0.46 20	0.00 1	0.00 1	0.17 14	0.02 14	2.73 35	0.54 45	0.00 1	0.00 1	0.00 1	0.00 1	0.01 10	0.01 21

PSMNet_ROB	0.41 22	0.54 54	1.69 32	0.00 1	0.00 1	0.92 50	0.00 1	0.27 5	0.53 24	0.00 1	0.00 1	0.25 22	0.03 24	3.43 45	0.47 41	0.00 1	0.00 1	0.00 1	0.00 1	0.05 36	0.03 45

RP	0.42 23	0.44 39	1.52 24	0.00 1	0.01 48	0.61 25	0.02 54	1.06 41	0.39 18	0.03 37	0.01 37	0.99 56	0.16 49	2.89 39	0.18 27	0.00 1	0.00 1	0.00 1	0.00 1	0.10 57	0.00 1

StereoDRNet	0.42 23	0.50 46	2.50 54	0.02 53	0.06 72	0.51 17	0.00 1	0.34 7	0.61 46	0.00 1	0.00 1	0.33 28	0.02 14	2.87 38	0.50 44	0.00 1	0.00 1	0.00 1	0.00 1	0.02 20	0.02 35

DRN-Test	0.43 25	0.13 12	1.95 39	0.03 58	0.09 79	0.61 25	0.00 1	1.43 58	0.61 46	0.00 1	0.00 1	0.18 17	0.04 31	3.32 44	0.11 21	0.00 1	0.00 1	0.00 1	0.00 1	0.01 10	0.01 21

NVstereo2D	0.45 26	0.11 9	1.91 38	0.00 1	0.00 1	0.59 24	0.00 1	1.39 56	0.61 46	0.03 37	0.00 1	0.33 28	0.00 1	3.94 53	0.07 16	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1

ccs	0.46 27	0.56 61	1.46 22	0.00 1	0.00 1	0.81 41	0.01 39	0.64 21	0.57 31	0.01 22	0.00 1	0.09 8	0.01 8	4.76 59	0.21 30	0.00 1	0.00 1	0.00 1	0.00 1	0.08 50	0.01 21

ccs_rob	0.46 27	0.54 54	1.68 30	0.00 1	0.00 1	0.98 54	0.01 39	0.70 25	0.57 31	0.00 1	0.00 1	0.07 5	0.01 8	4.35 57	0.17 26	0.00 1	0.00 1	0.00 1	0.00 1	0.11 62	0.03 45

CF-Net	0.47 29	0.53 52	2.05 44	0.00 1	0.00 1	0.41 11	0.01 39	1.18 50	0.40 19	0.13 51	0.05 55	1.30 66	0.45 60	2.43 32	0.30 35	0.00 1	0.00 1	0.00 1	0.00 1	0.09 53	0.00 1

TDLM	0.49 30	0.39 35	1.02 13	0.00 1	0.00 1	1.08 62	0.01 39	0.36 8	0.58 39	0.14 53	0.00 1	0.27 23	0.19 50	5.46 69	0.20 29	0.00 1	0.00 1	0.00 1	0.00 1	0.13 66	0.01 21

NCC-stereo	0.49 30	0.46 41	2.04 43	0.03 58	0.03 65	1.06 59	0.01 39	0.78 28	0.22 13	0.06 45	0.19 60	0.64 47	0.05 37	3.65 48	0.63 49	0.00 1	0.00 1	0.00 1	0.00 1	0.02 20	0.00 1

stereogan	0.49 30	0.22 17	2.39 51	0.00 1	0.00 1	0.63 29	0.02 54	0.79 29	0.72 68	0.06 45	0.00 1	1.68 72	0.46 61	2.63 34	0.04 10	0.00 1	0.00 1	0.00 1	0.00 1	0.14 68	0.00 1

RGC	0.50 33	0.53 52	2.42 52	0.00 1	0.01 48	0.65 30	0.00 1	1.07 42	0.17 8	0.01 22	0.62 67	1.00 57	0.19 50	3.03 41	0.28 34	0.00 1	0.00 1	0.00 1	0.00 1	0.05 36	0.02 35

SHD	0.50 33	0.51 51	3.61 73	0.05 65	0.00 1	0.01 1	0.01 39	0.72 26	0.61 46	1.57 75	0.00 1	0.34 31	0.09 41	1.79 21	0.49 43	0.00 1	0.00 1	0.00 1	0.01 76	0.09 53	0.03 45

PASM	0.51 35	0.32 22	3.07 64	0.00 1	0.00 1	0.31 7	0.00 1	0.46 12	0.12 4	0.03 37	0.00 1	0.44 38	0.02 14	5.22 66	0.04 10	0.00 1	0.00 1	0.00 1	0.00 1	0.12 63	0.03 45

G-Net	0.51 35	0.54 54	2.88 61	0.00 1	0.01 48	0.74 36	0.01 39	0.69 24	0.19 12	0.00 1	0.00 1	1.38 68	0.07 39	3.26 43	0.48 42	0.00 1	0.00 1	0.00 1	0.00 1	0.02 20	0.01 21

HSM-Net_RVC	0.52 37	0.06 3	1.45 21	0.00 1	0.00 1	0.61 25	0.00 1	3.83 89	0.52 22	0.14 53	0.00 1	0.22 21	0.03 24	3.50 47	0.04 10	0.01 79	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
CVANet_RVC	0.53 38	0.32 22	1.26 15	0.00 1	0.00 1	0.98 54	0.00 1	0.72 26	0.55 26	0.35 59	0.12 58	0.36 34	0.14 47	5.53 71	0.08 18	0.00 1	0.00 1	0.00 1	0.00 1	0.16 73	0.00 1

RYNet	0.53 38	0.19 15	1.79 35	0.00 1	0.00 1	0.55 20	0.00 1	1.12 47	0.61 46	0.02 31	0.00 1	0.83 51	0.04 31	5.05 64	0.45 40	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.02 35

GANetREF_RVC	0.54 40	0.70 74	2.06 45	0.00 1	0.00 1	1.14 66	0.00 1	1.61 64	0.58 39	0.00 1	0.01 37	0.29 25	0.04 31	4.21 55	0.06 14	0.00 1	0.00 1	0.00 1	0.00 1	0.03 28	0.02 35
Zhang, Feihu and Prisacariu, Victor and Yang, Ruigang and Torr, Philip HS: GA-Net: Guided Aggregation Net for End- to-end Stereo Matching. CVPR 2019
GANet	0.55 41	0.35 29	1.40 19	0.00 1	0.00 1	0.77 40	0.08 61	0.36 8	0.57 31	0.01 22	0.00 1	0.68 48	0.11 45	5.52 70	1.07 53	0.00 1	0.00 1	0.02 85	0.00 1	0.09 53	0.03 45

CFNet	0.56 42	0.50 46	1.58 26	0.00 1	0.00 1	1.06 59	0.01 39	1.08 43	0.57 31	0.00 1	0.01 37	0.12 10	0.01 8	5.87 79	0.19 28	0.00 1	0.00 1	0.00 1	0.00 1	0.13 66	0.02 35

PDISCO_ROB	0.56 42	0.57 62	2.07 46	0.17 81	0.00 1	1.56 74	0.01 39	2.27 78	0.51 21	0.08 47	0.01 37	0.74 49	0.28 57	2.37 30	0.13 24	0.00 1	0.00 1	0.00 1	0.00 1	0.33 92	0.00 1

PA-Net	0.57 44	0.50 46	2.73 60	0.03 58	0.06 72	0.89 47	0.02 54	0.66 22	0.63 63	0.00 1	0.00 1	0.43 37	0.07 39	4.82 60	0.55 47	0.01 79	0.00 1	0.00 1	0.00 1	0.02 20	0.04 56
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
CBMV	0.57 44	0.22 17	1.61 28	0.00 1	0.00 1	0.84 44	0.78 79	2.58 82	0.61 46	0.38 62	0.00 1	1.25 63	0.63 65	2.07 24	0.38 37	0.00 1	0.00 1	0.00 1	0.00 1	0.04 32	0.08 71
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
DANet	0.58 46	0.19 15	1.87 36	0.26 86	0.06 72	0.51 17	0.11 62	1.41 57	0.52 22	0.22 56	0.06 56	1.06 59	0.02 14	4.25 56	0.92 51	0.00 1	0.00 1	0.00 1	0.00 1	0.02 20	0.06 66

Anonymous Stereo	0.59 47	0.78 82	2.67 57	0.00 1	0.00 1	0.55 20	0.49 72	1.05 40	0.28 16	0.00 1	0.01 37	0.12 10	0.02 14	5.54 72	0.27 33	0.00 1	0.00 1	0.00 1	0.00 1	0.07 46	0.00 1

NOSS_ROB	0.61 48	0.17 14	0.63 1	0.19 83	0.00 1	0.67 32	0.00 1	2.38 79	0.57 31	0.13 51	0.00 1	0.94 53	0.00 1	6.27 84	0.10 20	0.00 1	0.00 1	0.00 1	0.00 1	0.04 32	0.10 77

PWC_ROB	0.62 49	0.54 54	3.22 66	0.00 1	0.00 1	0.22 4	0.00 1	2.18 76	0.61 46	0.00 1	0.20 61	0.29 25	0.03 24	3.46 46	1.22 55	0.00 1	0.00 1	0.00 1	0.00 1	0.39 93	0.01 21

HSM	0.63 50	0.36 32	0.89 7	0.00 1	0.00 1	0.52 19	0.01 39	0.98 38	0.56 27	0.00 1	0.01 37	2.27 78	0.65 66	6.18 81	0.11 21	0.00 1	0.00 1	0.00 1	0.00 1	0.01 10	0.00 1

XQC	0.68 51	0.72 78	3.82 75	0.06 68	0.00 1	0.39 10	0.01 39	1.92 70	0.61 46	0.03 37	0.01 37	0.46 40	0.21 52	2.84 37	2.26 66	0.00 1	0.00 1	0.00 1	0.01 76	0.15 70	0.05 60

RTSC	0.70 52	0.71 76	3.19 65	0.01 47	0.00 1	0.20 3	0.00 1	0.81 31	0.60 45	0.29 58	0.36 63	0.36 34	0.04 31	2.04 23	4.87 78	0.00 1	0.03 81	0.00 1	0.02 82	0.28 87	0.11 78

PVD	0.73 53	0.57 62	3.38 70	0.04 62	0.19 86	0.74 36	0.00 1	0.29 6	0.79 71	2.37 79	0.00 1	0.53 44	0.09 41	2.19 28	3.20 72	0.00 1	0.05 87	0.00 1	0.02 82	0.08 50	0.15 86

AANet_RVC	0.75 54	0.65 69	2.52 55	0.01 47	0.00 1	0.66 31	0.00 1	0.61 19	0.13 5	0.00 1	1.29 76	0.96 54	0.04 31	5.98 80	1.82 62	0.13 90	0.02 78	0.00 1	0.00 1	0.07 46	0.04 56

NaN_ROB	0.80 55	0.47 43	1.63 29	0.17 81	0.19 86	0.58 22	0.38 70	2.11 73	0.80 72	0.53 63	0.33 62	0.48 41	1.13 74	4.98 63	1.94 63	0.00 1	0.02 78	0.01 79	0.05 88	0.06 42	0.14 85

FBW_ROB	0.83 56	0.31 21	1.48 23	0.10 75	0.01 48	0.62 28	0.01 39	1.91 69	0.59 43	0.01 22	0.01 37	1.10 61	0.05 37	7.84 89	2.07 65	0.05 88	0.01 75	0.03 88	0.01 76	0.01 10	0.47 93

SGM-Forest	0.84 57	0.07 7	0.90 9	0.02 53	0.07 76	1.10 64	0.68 75	2.12 75	0.61 46	1.14 72	0.78 71	1.36 67	0.98 70	5.55 73	1.31 58	0.00 1	0.00 1	0.00 1	0.00 1	0.02 20	0.06 66
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
CBMV_ROB	0.84 57	0.06 3	1.73 33	0.00 1	0.00 1	1.37 70	0.47 71	2.57 81	0.58 39	2.13 77	0.76 70	1.06 59	1.38 76	4.04 54	0.54 45	0.00 1	0.00 1	0.00 1	0.00 1	0.03 28	0.05 60

SPS-STEREO	0.88 59	0.40 37	2.01 41	0.00 1	0.01 48	0.42 13	0.17 63	0.50 15	0.61 46	2.19 78	0.79 72	2.04 77	0.27 56	5.10 65	2.89 71	0.00 1	0.00 1	0.00 1	0.00 1	0.03 28	0.11 78
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
ADCRef	0.92 60	0.48 44	3.35 68	0.02 53	0.02 55	0.90 48	0.32 68	0.61 19	0.65 65	0.72 64	0.17 59	0.17 14	0.82 69	2.15 26	7.95 81	0.00 1	0.00 1	0.00 1	0.00 1	0.07 46	0.03 45

PWCDC_ROB	0.94 61	0.91 86	2.52 55	0.00 1	0.00 1	1.15 67	0.00 1	2.11 73	1.18 79	0.00 1	8.99 89	0.35 32	0.09 41	0.64 10	0.40 38	0.00 1	0.00 1	0.00 1	0.00 1	0.50 94	0.03 45

SAMSARA	1.01 62	0.70 74	3.29 67	0.20 84	0.06 72	1.65 75	0.05 60	0.48 14	0.53 24	2.91 83	0.68 69	2.97 79	1.08 72	3.89 52	1.54 59	0.00 1	0.04 83	0.00 1	0.02 82	0.05 36	0.03 45

CSAN	1.05 63	0.61 66	2.46 53	0.09 71	0.08 78	0.71 35	0.71 77	2.20 77	0.93 75	0.85 66	2.92 79	0.79 50	1.46 77	5.37 67	1.58 60	0.01 79	0.01 75	0.00 1	0.00 1	0.09 53	0.05 60

NVStereoNet_ROB	1.08 64	0.39 35	2.15 49	0.08 70	0.02 55	0.45 15	0.02 54	0.89 34	0.67 66	0.10 50	5.85 83	1.90 74	1.11 73	6.20 82	1.14 54	0.20 92	0.00 1	0.03 88	0.07 89	0.12 63	0.13 83
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
DispFullNet	1.20 65	0.63 67	3.58 72	1.00 93	0.82 92	0.41 11	0.00 1	1.44 59	0.61 46	2.55 80	0.64 68	1.10 61	8.95 93	0.84 11	1.22 55	0.00 1	0.02 78	0.00 1	0.00 1	0.24 83	0.02 35

DPSNet	1.25 66	0.28 20	5.34 82	0.28 88	0.07 76	1.50 71	0.00 1	3.53 88	0.69 67	0.03 37	0.00 1	0.45 39	2.10 81	5.67 75	4.61 76	0.02 84	0.04 83	0.00 1	0.02 82	0.23 82	0.09 74

DeepPrunerF	1.33 67	0.59 64	8.62 90	0.02 53	0.09 79	0.90 48	0.00 1	1.34 53	6.72 93	0.22 56	0.01 37	0.21 19	0.14 47	5.64 74	1.99 64	0.00 1	0.00 1	0.01 79	0.02 82	0.14 68	0.02 35

ADCP+	1.36 68	0.76 81	6.73 85	0.01 47	0.02 55	1.32 69	0.80 80	0.81 31	0.72 68	0.01 22	0.01 37	0.96 54	1.05 71	0.33 9	13.50 94	0.00 1	0.00 1	0.00 1	0.00 1	0.06 42	0.02 35

ADCMid	1.38 69	0.90 85	6.15 84	0.04 62	0.03 65	1.67 76	0.18 64	1.13 48	0.93 75	0.89 67	0.58 66	0.86 52	0.71 67	1.03 12	11.96 89	0.01 79	0.00 1	0.29 96	0.08 91	0.17 76	0.03 45

AnyNet_C32	1.44 70	1.47 90	6.09 83	0.12 76	0.19 86	1.74 77	1.11 83	0.84 33	0.56 27	1.30 73	0.46 65	0.21 19	0.59 63	2.74 36	11.09 87	0.02 84	0.00 1	0.06 90	0.00 1	0.16 73	0.05 60

ADCPNet	1.44 70	0.55 59	7.88 88	0.01 47	0.19 86	2.78 82	0.27 66	0.90 35	0.38 17	0.04 43	0.02 52	1.43 70	1.81 80	2.22 29	9.08 82	0.00 1	0.47 95	0.00 1	0.64 96	0.04 32	0.15 86

ADCL	1.52 72	0.44 39	4.70 79	0.07 69	0.03 65	2.21 79	2.35 87	0.92 36	0.80 72	1.57 75	1.11 74	1.25 63	1.55 78	1.77 20	11.53 88	0.01 79	0.00 1	0.00 1	0.00 1	0.10 57	0.01 21

MDST_ROB	1.54 73	0.06 3	3.42 71	0.14 77	0.16 81	5.42 92	0.32 68	4.33 91	0.61 46	6.08 88	1.25 75	1.04 58	1.16 75	5.81 78	0.83 50	0.00 1	0.00 1	0.00 1	0.00 1	0.01 10	0.11 78

SGM_RVC	1.54 73	0.35 29	1.60 27	0.15 78	0.17 83	1.25 68	0.77 78	1.97 72	0.88 74	4.31 84	2.46 78	4.36 83	2.37 82	6.98 87	1.78 61	0.21 93	0.23 94	0.20 95	0.23 94	0.26 85	0.31 90
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
ADCS	1.72 75	1.22 89	8.11 89	0.00 1	0.01 48	0.88 46	0.68 75	1.44 59	0.63 63	2.60 81	0.11 57	1.42 69	0.26 55	1.97 22	14.92 96	0.00 1	0.00 1	0.00 1	0.00 1	0.08 50	0.07 70

SANet	1.99 76	0.71 76	3.37 69	0.03 58	0.03 65	2.85 83	0.85 81	2.42 80	8.33 95	1.55 74	4.13 81	4.29 82	3.23 84	5.39 68	2.47 68	0.00 1	0.00 1	0.00 1	0.00 1	0.06 42	0.06 66

WCMA_ROB	2.01 77	0.34 24	1.89 37	0.04 62	0.00 1	1.07 61	0.60 73	1.09 44	1.21 80	5.12 86	9.87 92	8.09 90	4.48 86	3.74 50	2.55 69	0.00 1	0.00 1	0.01 79	0.00 1	0.02 20	0.13 83

MSMD_ROB	2.04 78	0.40 37	2.13 48	0.00 1	0.00 1	4.88 91	0.22 65	0.93 37	1.82 86	5.71 87	7.79 88	7.39 89	3.72 85	4.62 58	1.22 55	0.00 1	0.00 1	0.00 1	0.00 1	0.00 1	0.02 35

pmcnn	2.10 79	0.59 64	4.52 78	0.27 87	0.86 93	0.34 9	0.29 67	1.69 67	0.75 70	0.36 60	7.50 87	14.39 96	0.56 62	4.90 61	4.69 77	0.00 1	0.00 1	0.00 1	0.00 1	0.16 73	0.04 56

Abc-Net	2.21 80	0.84 83	4.83 80	0.09 71	0.00 1	4.19 90	4.09 90	6.19 93	1.76 85	0.95 68	1.86 77	5.20 84	4.77 87	6.26 83	2.72 70	0.00 1	0.00 1	0.00 1	0.00 1	0.31 91	0.22 89

AnyNet_C01	2.33 81	2.14 95	18.59 95	0.15 78	0.18 84	2.43 81	2.34 86	1.29 52	1.24 81	1.03 69	0.83 73	1.25 63	0.78 68	3.75 51	10.06 84	0.05 88	0.00 1	0.07 92	0.00 1	0.30 90	0.12 81

SGM+DAISY	2.38 82	0.98 87	4.37 77	0.48 89	0.33 90	2.24 80	1.94 85	1.16 49	1.47 83	6.62 89	5.89 84	5.77 86	5.09 88	5.76 77	4.39 75	0.16 91	0.15 92	0.06 90	0.14 92	0.26 85	0.40 92

PWCK	2.44 83	1.65 91	7.31 87	0.22 85	0.02 55	3.12 84	4.74 93	3.06 85	3.48 89	0.72 64	4.42 82	3.40 80	1.67 79	9.06 91	3.36 73	0.39 95	0.01 75	0.00 1	0.00 1	2.19 96	0.08 71

MeshStereo	2.61 84	0.65 69	2.71 59	0.02 53	0.02 55	1.55 73	0.94 82	3.02 84	1.29 82	11.21 92	7.34 86	11.54 93	2.74 83	6.34 85	2.43 67	0.00 1	0.00 1	0.00 1	0.00 1	0.28 87	0.03 45
C. Zhang, Z. Li, Y. Cheng, R. Cai, H. Chao, Y. Rui: MeshStereo: A Global Stereo Model with Mesh Alignment Regularization for View Interpolation. ICCV 2015
ELAS	2.75 85	0.35 29	2.67 57	0.05 65	0.04 69	6.72 93	1.83 84	2.84 83	3.38 88	8.25 91	9.14 90	3.95 81	6.15 89	3.73 49	5.70 80	0.00 1	0.04 83	0.01 79	0.01 76	0.10 57	0.08 71
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ELAS_RVC	2.84 86	0.34 24	3.86 76	0.05 65	0.05 71	1.52 72	5.02 94	1.93 71	5.28 90	7.10 90	6.96 85	6.26 88	8.34 91	4.96 62	4.90 79	0.00 1	0.04 83	0.01 79	0.01 76	0.10 57	0.09 74
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
Nwc_Net	2.87 87	0.63 67	5.17 81	0.00 1	0.04 69	3.84 87	4.22 91	8.06 94	0.93 75	0.37 61	0.44 64	5.87 87	6.66 90	7.87 90	12.93 91	0.00 1	0.00 1	0.00 1	0.00 1	0.24 83	0.12 81

FC-DCNN	3.42 88	0.11 9	2.03 42	0.00 1	0.00 1	3.18 85	0.60 73	3.13 86	2.98 87	11.40 93	12.28 93	13.93 94	8.68 92	5.70 76	4.27 74	0.00 1	0.00 1	0.00 1	0.00 1	0.01 10	0.06 66

RTS	4.10 89	1.68 92	43.40 96	0.09 71	0.02 55	3.95 88	0.00 1	1.61 64	0.61 46	1.09 70	0.01 37	1.97 75	0.22 53	16.87 93	10.15 85	0.00 1	0.09 89	0.00 1	0.00 1	0.20 78	0.05 60

RTSA	4.10 89	1.68 92	43.40 96	0.09 71	0.02 55	3.95 88	0.00 1	1.61 64	0.61 46	1.09 70	0.01 37	1.97 75	0.22 53	16.87 93	10.15 85	0.00 1	0.09 89	0.00 1	0.00 1	0.20 78	0.05 60

LSM	4.58 91	0.88 84	17.79 94	3.73 97	38.79 98	0.85 45	2.37 88	0.54 16	1.64 84	2.88 82	3.09 80	0.62 45	0.60 64	3.16 42	1.06 52	0.00 1	0.00 1	0.00 1	0.00 1	0.12 63	13.55 98

MADNet+	4.65 92	2.70 96	46.68 98	0.01 47	0.02 55	2.06 78	0.03 58	3.17 87	1.07 78	0.02 31	0.01 37	1.84 73	0.31 58	21.62 97	13.04 92	0.00 1	0.03 81	0.00 1	0.00 1	0.20 78	0.19 88

edge stereo	7.09 93	3.32 97	14.31 93	0.96 92	1.13 94	26.20 98	16.27 98	13.37 97	6.61 92	4.33 85	17.00 95	5.35 85	12.69 95	7.34 88	9.08 82	0.61 96	0.19 93	0.02 85	0.04 87	1.40 95	1.58 95

DPSimNet_ROB	7.90 94	5.46 98	13.55 92	3.66 96	5.16 97	7.26 94	9.60 97	6.11 92	7.46 94	13.33 94	9.50 91	11.03 91	11.89 94	10.08 92	12.35 90	4.36 97	5.55 97	5.35 98	5.34 98	4.96 97	5.98 97

SGM-ForestM	8.89 95	0.68 72	3.61 73	0.57 91	0.18 84	14.89 97	5.99 95	9.62 96	10.59 96	20.43 96	17.16 96	39.22 98	20.29 97	20.63 96	13.19 93	0.04 86	0.14 91	0.00 1	0.07 89	0.15 70	0.36 91

MADNet++	9.87 96	2.12 94	7.11 86	7.99 98	4.58 96	14.38 96	6.90 96	9.23 95	5.88 91	16.88 95	17.47 97	11.07 92	15.72 96	24.26 98	19.76 97	8.52 98	6.39 98	4.85 97	2.69 97	5.96 98	5.73 96

LE_ROB	9.96 97	0.54 54	9.62 91	1.40 95	1.24 95	3.29 86	4.64 92	4.31 90	43.01 98	51.54 98	13.34 94	14.15 95	24.60 99	6.66 86	20.28 98	0.04 86	0.05 87	0.12 94	0.20 93	0.07 46	0.09 74

MANE	9.96 97	0.69 73	2.88 61	1.22 94	0.61 91	13.15 95	3.46 89	15.62 98	22.27 97	22.74 97	30.00 98	29.74 97	23.22 98	17.17 95	13.78 95	0.29 94	0.56 96	0.09 93	0.61 95	0.15 70	1.00 94

DGTPSM_ROB	81.49 99	60.74 99	99.95 101	61.20 99	86.99 101	99.11 101	100.00 101	95.92 100	99.99 101	73.08 99	100.00 100	70.81 99	100.00 101	77.39 99	99.98 101	35.10 99	97.48 102	35.12 99	99.89 102	46.16 99	90.85 99

DPSMNet_ROB	81.54 100	60.77 100	99.95 101	61.42 100	87.05 102	99.11 101	100.00 101	95.93 101	99.99 101	73.11 100	100.00 100	70.84 100	100.00 101	77.43 100	99.98 101	35.39 100	97.54 103	35.21 100	99.89 102	46.30 100	90.94 100

MEDIAN_ROB	93.37 101	98.75 103	96.05 100	90.90 103	90.53 103	85.63 99	76.45 99	92.03 99	89.53 99	95.05 101	94.99 99	96.02 101	93.85 100	90.36 101	87.21 99	98.36 104	95.79 99	99.61 103	98.77 101	98.80 104	98.72 104

DPSM	94.00 102	79.32 101	100.00 103	83.19 101	85.69 99	100.00 103	100.00 101	100.00 102	100.00 103	100.00 102	100.00 100	100.00 102	100.00 101	100.00 103	100.00 103	89.53 101	97.42 100	74.97 101	95.13 99	83.79 101	90.98 101

DPSM_ROB	94.00 102	79.32 101	100.00 103	83.19 101	85.69 99	100.00 103	100.00 101	100.00 102	100.00 103	100.00 102	100.00 100	100.00 102	100.00 101	100.00 103	100.00 103	89.53 101	97.42 100	74.97 101	95.13 99	83.79 101	90.98 101

AVERAGE_ROB	98.45 104	99.13 104	95.46 99	100.00 105	100.00 105	93.36 100	81.95 100	100.00 102	99.97 100	100.00 102	100.00 100	100.00 102	100.00 101	99.72 102	99.40 100	100.00 105	100.00 105	100.00 104	100.00 104	100.00 105	100.00 105

LSM0	99.11 105	99.97 105	100.00 103	95.93 104	98.59 104	100.00 103	100.00 101	100.00 102	100.00 103	100.00 102	100.00 100	100.00 102	100.00 101	100.00 103	100.00 103	93.89 103	99.97 104	100.00 104	100.00 104	96.64 103	97.25 103

MSMDNet													0.34 59

Low-res two-view benchmark