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

PMTNet	5.81 1	1.06 1	4.24 1	2.87 1	0.77 1	9.07 5	7.38 33	14.08 1	24.63 15	9.23 1	11.01 1	7.87 1	5.43 2	12.69 5	5.77 4	0.01 2	0.02 5	0.00 1	0.00 1	0.16 6	0.01 1

DPM-Stereo	6.74 2	1.93 11	10.02 21	4.28 2	1.64 6	13.54 13	1.52 2	15.16 2	6.63 1	14.30 3	26.81 13	10.90 3	5.19 1	14.78 11	7.56 7	0.00 1	0.06 18	0.05 11	0.00 1	0.17 7	0.17 10

R-Stereo Train	7.04 3	1.34 4	7.89 11	6.23 16	2.30 11	7.51 2	5.54 15	23.96 9	6.94 2	16.26 9	33.57 39	12.04 4	5.61 3	7.67 1	3.74 1	0.01 2	0.05 12	0.01 5	0.00 1	0.10 1	0.12 4

R-Stereo	7.04 3	1.34 4	7.89 11	6.23 16	2.30 11	7.51 2	5.54 15	23.96 9	6.94 2	16.26 9	33.57 39	12.04 4	5.61 3	7.67 1	3.74 1	0.01 2	0.05 12	0.01 5	0.00 1	0.10 1	0.12 4

HITNet	7.83 5	2.93 24	8.39 14	4.76 3	1.48 5	12.75 9	4.64 10	20.97 3	14.37 4	15.14 6	22.15 3	14.57 7	10.62 9	14.38 9	8.14 9	0.04 6	0.01 2	0.48 35	0.02 14	0.57 16	0.10 3
Vladimir Tankovich, Christian Häne, Yinda Zhang, Adarsh Kowdle, Sean Fanello, Sofien Bouaziz: HITNet: Hierarchical Iterative Tile Refinement Network for Real-time Stereo Matching. CVPR 2021
DN-CSS_ROB	8.17 6	2.59 19	14.14 35	6.11 14	3.22 21	6.31 1	3.53 4	22.92 7	20.59 9	16.66 12	30.05 22	9.31 2	7.95 7	13.49 7	4.78 3	0.72 41	0.00 1	0.45 33	0.00 1	0.54 14	0.09 2

MLCV	9.32 7	2.22 16	12.55 30	5.22 5	1.32 2	13.79 14	1.35 1	21.14 4	23.65 14	21.91 27	31.37 27	17.14 9	7.57 6	16.00 18	10.18 17	0.22 16	0.01 2	0.03 9	0.02 14	0.40 9	0.29 14

BEATNet_4x	9.37 8	4.85 48	12.05 28	5.48 6	1.35 4	13.15 10	6.15 22	25.06 12	16.74 5	16.18 8	24.65 5	17.29 10	12.51 14	16.91 23	10.92 22	0.38 25	0.18 29	0.81 47	0.05 19	1.95 47	0.65 23

ccs	9.74 9	1.25 3	7.47 9	5.77 9	2.16 9	7.74 4	3.54 5	33.83 49	35.91 45	13.42 2	23.80 4	22.01 16	11.97 12	11.51 3	9.09 14	0.41 29	1.75 76	0.17 18	0.17 28	0.63 18	2.12 57

CFNet_RVC	9.87 10	2.18 15	5.49 3	7.45 27	5.25 42	15.76 25	5.98 18	21.89 5	21.87 11	14.96 4	30.02 21	25.68 32	13.68 17	12.46 4	10.94 23	0.08 9	0.06 18	1.83 64	0.22 30	1.09 29	0.43 18

AdaStereo	10.22 11	3.63 33	9.14 19	9.15 41	3.24 22	15.79 26	5.39 14	31.94 38	25.42 16	18.39 15	26.47 9	19.44 13	9.50 8	16.60 22	8.59 11	0.44 32	0.05 12	0.40 31	0.00 1	0.57 16	0.16 9
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021.
iResNet	10.26 12	3.10 29	15.72 44	7.35 25	2.13 8	13.86 15	7.07 31	22.80 6	28.01 24	20.10 20	30.40 24	16.79 8	11.06 10	14.46 10	10.29 20	0.39 28	0.05 12	0.00 1	0.07 24	0.73 21	0.73 24

ccs_rob	10.52 13	2.12 13	7.47 9	6.02 13	3.07 18	19.32 37	4.16 6	25.06 12	27.77 23	20.21 21	27.93 18	30.56 42	13.32 15	14.95 13	7.25 6	0.04 6	0.05 12	0.21 19	0.05 19	0.55 15	0.27 13

CFNet	10.67 14	2.33 17	8.90 16	6.66 22	4.09 28	16.05 27	4.26 8	30.44 32	31.01 33	18.53 16	26.48 10	23.69 24	13.68 17	15.27 15	10.59 21	0.05 8	0.02 5	0.44 32	0.05 19	0.73 21	0.23 11

DeepPruner_ROB	10.77 15	4.56 43	13.13 31	6.37 19	4.28 31	10.14 6	6.86 27	36.42 65	19.59 8	17.87 13	27.37 16	23.19 20	12.26 13	19.09 30	10.23 19	1.05 54	0.48 54	0.23 21	0.15 27	0.89 27	1.17 36

HSM-Net_RVC	10.88 16	1.23 2	5.46 2	5.57 7	2.63 14	20.12 39	6.06 20	27.67 19	28.35 25	20.63 23	25.51 6	34.01 54	14.38 25	16.18 20	9.29 15	0.11 10	0.07 21	0.01 5	0.00 1	0.13 3	0.14 8
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
NOSS_ROB	10.99 17	3.81 37	6.88 6	6.69 23	2.30 11	16.06 28	10.94 45	30.06 30	32.41 36	16.10 7	18.72 2	22.78 18	14.21 22	17.52 25	8.62 12	2.35 71	1.84 78	3.32 83	1.57 70	2.06 51	1.52 44

DMCA	11.27 18	2.78 20	10.03 22	10.86 57	5.57 46	11.86 8	6.42 23	23.55 8	19.24 6	19.96 18	32.17 34	25.00 30	27.88 68	13.03 6	14.75 30	0.30 22	0.30 37	0.29 24	0.23 32	0.76 23	0.33 16

HSM	11.33 19	1.80 8	7.04 8	6.13 15	3.99 27	16.16 30	6.92 29	29.69 28	19.39 7	22.95 30	26.05 8	41.08 76	15.79 31	20.25 35	8.89 13	0.02 5	0.03 8	0.01 5	0.00 1	0.15 4	0.34 17

RASNet	11.59 20	1.71 7	10.36 23	6.30 18	5.66 48	15.48 24	6.12 21	34.50 55	20.86 10	16.61 11	32.51 35	22.04 17	23.00 51	19.44 31	16.15 37	0.71 39	0.03 8	0.05 11	0.01 12	0.15 4	0.13 7

iResNet_ROB	11.71 21	1.96 12	10.77 24	6.01 12	3.68 24	15.13 22	2.34 3	31.72 37	37.58 52	26.16 44	32.77 36	25.88 33	15.34 28	16.35 21	7.94 8	0.19 14	0.01 2	0.00 1	0.00 1	0.31 8	0.12 4

FADNet-RVC-Resample	11.72 22	4.35 41	30.33 75	9.46 45	3.90 26	14.11 16	7.17 32	25.78 14	23.31 13	23.27 33	29.52 20	19.50 14	15.00 27	15.38 16	8.33 10	0.53 35	0.50 55	0.49 36	0.50 47	1.38 39	1.67 48

CBMV_ROB	11.77 23	2.89 22	6.92 7	5.20 4	2.88 17	14.12 17	4.79 11	26.69 16	29.84 27	26.17 45	31.92 31	23.61 23	20.74 47	20.20 34	10.18 17	1.78 65	1.80 77	2.17 69	1.26 66	1.61 41	0.55 21

FADNet_RVC	11.94 24	5.17 52	39.74 87	6.44 20	3.46 23	11.42 7	4.17 7	29.16 26	26.89 20	17.97 14	25.96 7	13.41 6	15.67 29	15.84 17	13.54 27	0.90 48	0.75 64	0.55 39	1.20 64	3.65 68	2.85 64

iResNetv2_ROB	12.00 25	2.91 23	13.46 32	5.82 10	3.73 25	13.15 10	6.45 24	34.15 50	36.02 47	25.25 42	33.81 42	29.79 40	14.24 23	13.61 8	6.10 5	0.21 15	0.02 5	0.12 14	0.00 1	0.82 25	0.25 12

NLCA_NET_v2_RVC	12.05 26	3.43 31	16.25 49	7.62 30	5.58 47	14.57 21	6.00 19	32.98 46	27.41 21	21.69 26	31.40 28	25.21 31	13.47 16	14.96 14	15.60 33	0.84 43	0.33 41	0.35 27	0.35 40	1.29 36	1.78 50
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
CC-Net-ROB	12.13 27	3.51 32	15.82 45	7.52 28	5.50 45	14.30 19	6.60 26	32.79 44	27.59 22	21.97 28	32.14 33	24.76 29	14.02 19	14.82 12	16.06 34	0.85 45	0.35 43	0.34 26	0.35 40	1.30 38	1.94 53

SGM-Forest	12.92 28	1.87 9	6.61 4	5.68 8	2.05 7	23.19 52	11.30 48	34.24 51	30.77 32	26.54 47	31.99 32	29.98 41	15.74 30	21.17 39	13.36 26	1.05 54	0.33 41	0.84 48	0.01 12	0.71 20	0.97 29
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
STTStereo	12.97 29	4.88 49	26.57 69	7.54 29	5.41 44	13.43 12	6.49 25	30.53 33	26.35 19	25.14 41	31.43 29	32.00 45	14.12 20	16.01 19	14.83 31	0.29 21	0.31 39	0.67 43	1.08 61	1.21 33	1.11 33

TDLM	12.98 30	3.78 36	9.77 20	10.75 54	4.94 38	16.10 29	14.08 55	35.05 58	31.87 35	22.36 29	26.51 12	26.09 34	14.97 26	24.73 49	13.11 25	1.05 54	0.05 12	1.21 60	0.27 35	1.83 44	1.00 31

AANet_RVC	13.16 31	5.34 53	10.83 26	8.20 38	4.44 32	14.13 18	9.46 38	28.78 25	37.67 53	23.44 34	37.47 51	23.48 21	15.83 32	22.39 41	16.63 38	1.67 63	0.86 67	0.24 22	0.02 14	0.63 18	1.60 46

CVANet_RVC	13.24 32	3.39 30	8.43 15	8.42 39	5.04 40	18.52 34	11.72 50	32.03 40	33.85 39	23.17 32	32.77 36	29.78 39	16.51 33	23.20 42	11.28 24	0.94 49	0.04 10	1.19 59	0.47 45	3.13 63	1.01 32

FADNet-RVC	13.27 33	11.57 86	39.71 86	7.94 36	4.50 33	15.41 23	6.95 30	27.96 20	22.98 12	20.37 22	30.73 25	26.86 35	11.31 11	20.50 37	14.28 29	0.14 11	0.14 28	0.13 16	0.18 29	2.83 58	0.86 26

DLCB_ROB	13.35 34	3.00 26	9.12 17	9.43 44	5.68 49	21.80 46	10.12 40	29.19 27	29.92 28	27.71 50	31.45 30	32.37 46	19.50 42	19.02 29	16.73 39	0.22 16	0.04 10	0.38 30	0.06 23	0.49 12	0.85 25

StereoDRNet-Refined	13.36 35	2.80 21	10.82 25	7.94 36	3.10 19	18.95 35	4.45 9	28.47 24	29.65 26	29.11 54	41.47 68	24.44 27	17.64 38	23.34 43	21.66 44	0.16 12	0.07 21	0.53 38	0.32 38	0.77 24	1.64 47
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
CBMV	13.69 36	3.63 33	8.02 13	5.93 11	2.69 15	22.57 49	12.44 52	29.81 29	31.57 34	31.20 63	33.79 41	31.04 43	17.41 37	25.28 52	14.11 28	0.71 39	0.60 59	0.60 40	0.11 25	1.11 31	1.13 35
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
FADNet	14.18 37	10.92 83	37.56 82	7.79 34	6.64 55	17.48 31	6.91 28	34.73 56	34.03 40	18.85 17	27.02 14	28.26 36	14.24 23	17.77 27	10.09 16	0.41 29	0.67 62	0.37 28	0.37 42	8.13 89	1.34 38

pmcnn	14.59 38	3.68 35	19.78 60	6.83 24	4.51 34	21.27 44	14.67 58	27.37 17	30.54 29	27.09 49	40.19 63	38.71 67	18.53 41	17.70 26	19.27 40	0.84 43	0.09 26	0.00 1	0.00 1	0.48 10	0.31 15

NVstereo2D	15.15 39	2.97 25	16.28 50	9.23 42	7.56 64	30.03 68	9.53 39	42.80 89	42.72 65	15.11 5	27.03 15	23.16 19	16.93 35	24.28 46	15.45 32	3.07 79	0.43 50	1.44 61	0.61 50	6.74 83	7.60 92

StereoDRNet	15.30 40	4.47 42	14.35 39	10.71 53	8.69 70	25.02 57	11.03 46	39.66 78	35.62 44	29.17 55	28.81 19	34.05 55	17.80 39	18.80 28	23.77 52	0.41 29	0.30 37	0.15 17	0.14 26	1.64 42	1.42 41

DRN-Test	15.88 41	3.83 38	14.70 40	9.96 49	8.02 68	28.93 67	14.65 57	42.34 85	38.25 56	28.18 53	34.31 45	29.22 38	18.35 40	20.01 33	22.68 48	0.37 24	0.41 46	0.46 34	0.28 36	1.28 35	1.34 38

PA-Net	16.29 42	6.88 65	27.34 70	9.88 47	11.12 82	21.77 45	28.15 95	32.06 41	39.49 57	19.99 19	27.91 17	23.50 22	20.59 46	20.25 35	24.04 55	0.24 18	2.26 85	0.22 20	4.59 91	1.13 32	4.48 81
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
DISCO	16.36 43	1.88 10	16.15 47	7.78 33	4.19 29	26.92 61	10.21 42	30.31 31	44.26 72	21.35 25	33.91 43	35.80 62	22.40 50	35.65 84	33.87 83	0.18 13	0.08 23	0.04 10	0.04 18	1.71 43	0.54 20

NaN_ROB	16.51 44	6.51 61	16.22 48	10.53 52	4.64 36	31.76 74	17.78 71	37.00 69	43.37 68	29.68 59	31.29 26	32.67 50	20.32 45	28.00 61	16.06 34	0.38 25	0.41 46	0.32 25	0.41 43	0.87 26	1.94 53

DANet	16.85 45	8.00 68	26.04 68	14.56 75	7.25 62	20.59 40	5.37 13	26.22 15	26.02 18	32.17 68	36.00 46	37.20 64	29.77 74	30.75 65	24.62 60	1.14 58	1.34 71	2.21 70	0.51 48	3.18 64	4.07 77

PSMNet_ROB	16.90 46	5.45 54	14.16 36	13.77 69	7.25 62	27.65 63	32.74 99	43.21 91	37.88 54	24.00 35	32.97 38	34.55 58	16.69 34	17.41 24	23.96 54	0.38 25	0.22 32	0.93 50	2.05 77	1.83 44	0.96 28

GANet	16.92 47	4.32 40	12.46 29	10.84 56	4.24 30	23.04 51	15.36 62	39.91 79	34.47 42	32.42 69	45.10 83	40.04 74	27.54 65	23.61 45	20.20 41	1.02 53	0.08 23	0.67 43	0.22 30	1.87 46	0.97 29

NCCL2	17.23 48	6.37 60	14.30 38	23.45 97	7.18 61	24.16 55	17.05 69	34.39 53	25.88 17	31.55 64	38.01 54	42.09 81	26.81 61	20.94 38	22.69 49	0.45 33	0.23 33	2.63 75	1.80 73	2.02 50	2.57 61

ADCRef	17.30 49	6.73 63	42.01 90	10.11 50	8.78 71	26.61 60	10.45 43	31.29 35	32.75 38	31.19 62	41.62 70	18.47 12	19.78 44	24.29 47	34.06 85	0.76 42	0.42 49	2.16 68	1.79 72	1.23 34	1.56 45

XPNet_ROB	17.33 50	4.84 47	15.44 43	11.15 60	6.64 55	20.88 41	16.10 64	36.26 64	39.54 58	31.80 67	41.32 67	41.30 78	23.91 52	24.67 48	27.40 73	1.07 57	0.74 63	0.79 46	0.25 33	1.10 30	1.31 37

RP	17.58 51	4.92 50	15.27 42	15.44 80	11.95 88	21.05 42	11.75 51	27.64 18	45.09 75	22.99 31	42.05 72	39.39 70	27.74 67	25.54 53	22.33 47	5.45 93	0.59 58	4.33 86	1.62 71	3.53 67	2.86 65

SuperB	17.60 52	6.22 59	56.40 101	7.91 35	5.07 41	19.70 38	9.03 36	24.87 11	50.15 86	26.25 46	39.13 58	17.61 11	20.84 48	22.28 40	25.31 65	0.70 38	0.31 39	1.10 55	0.71 53	16.58 103	1.86 51

ETE_ROB	17.67 53	9.48 75	17.66 53	13.60 67	4.57 35	23.03 50	21.09 80	32.49 43	35.24 43	30.49 60	38.06 55	46.18 90	24.49 54	26.40 55	24.21 57	0.46 34	0.24 34	1.14 56	0.84 57	1.52 40	2.13 58

PWCDC_ROB	17.80 54	9.62 76	18.38 57	17.95 86	7.74 65	24.61 56	5.56 17	34.41 54	52.79 90	29.66 58	50.20 91	20.63 15	19.69 43	29.85 64	20.87 42	6.18 94	0.26 36	0.10 13	0.05 19	5.36 81	2.00 55

MDST_ROB	17.92 55	1.60 6	13.68 33	13.88 70	6.44 53	43.05 92	14.86 59	42.74 88	41.66 63	43.25 90	42.85 75	28.43 37	17.28 36	29.35 63	16.06 34	0.88 46	0.11 27	0.63 42	0.47 45	0.50 13	0.64 22

Nwc_Net	17.98 56	4.67 45	17.79 56	14.26 71	11.65 87	25.14 58	14.88 60	40.85 82	39.83 59	21.07 24	43.24 78	34.50 57	27.64 66	26.44 56	25.03 61	3.22 80	0.18 29	2.52 74	3.18 82	1.98 48	1.51 43

ADCP+	18.16 57	4.61 44	32.94 78	9.31 43	10.23 78	28.87 66	11.69 49	33.43 47	36.49 48	29.28 57	40.49 64	24.18 25	24.61 56	33.89 75	35.33 86	0.25 19	0.40 45	2.00 65	1.17 63	2.16 52	1.87 52

PWC_ROB	18.37 58	10.55 80	25.19 67	10.21 51	6.29 52	23.75 53	4.79 11	35.93 63	48.75 79	32.56 70	44.46 82	34.68 59	24.78 57	30.99 66	25.72 69	1.38 61	0.08 23	1.57 62	0.26 34	2.18 53	3.17 69

AF-Net	18.38 59	5.60 55	14.21 37	15.95 81	10.78 80	22.26 47	10.45 43	35.39 60	50.22 87	24.71 40	37.77 52	41.55 79	29.24 73	29.11 62	26.67 70	4.35 88	0.06 18	4.56 87	0.90 58	2.42 54	1.38 40

Anonymous Stereo	18.60 60	11.75 87	49.81 96	14.43 73	12.02 90	14.33 20	23.23 89	32.16 42	43.08 67	24.32 38	34.16 44	24.24 26	14.12 20	31.67 69	30.84 80	0.89 47	0.91 68	1.74 63	1.92 76	3.23 65	3.23 70

stereogan	18.62 61	3.07 27	16.31 51	13.04 66	9.99 76	35.74 84	9.09 37	38.17 76	45.27 76	24.39 39	41.24 66	39.93 72	25.46 59	31.29 68	24.33 59	1.80 66	0.91 68	2.71 77	0.75 56	5.15 80	3.70 71

edge stereo	18.75 62	5.12 51	17.66 53	10.77 55	7.84 66	22.41 48	11.03 46	35.23 59	42.07 64	34.62 74	42.74 74	41.28 77	35.50 86	25.14 51	24.17 56	2.42 75	2.42 87	6.35 91	1.45 68	2.89 61	3.80 72

RYNet	18.92 63	4.69 46	16.88 52	10.99 58	15.84 96	46.72 94	16.57 66	37.84 73	48.03 77	26.92 48	26.48 10	37.84 66	24.92 58	19.80 32	31.04 81	0.25 19	0.25 35	0.62 41	0.03 17	6.34 82	6.39 87

LALA_ROB	19.19 64	7.15 67	15.93 46	12.96 65	5.30 43	28.30 65	23.21 88	42.51 86	36.69 49	33.45 73	40.81 65	50.96 95	24.53 55	27.92 58	25.66 68	0.60 36	0.44 51	2.04 66	1.22 65	2.54 56	1.47 42

SGM_RVC	19.52 65	2.12 13	6.79 5	6.59 21	1.33 3	38.20 89	15.74 63	38.13 75	34.07 41	45.71 94	41.91 71	51.41 96	38.10 90	38.77 91	27.80 76	0.61 37	0.36 44	0.51 37	0.33 39	1.03 28	0.90 27
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
RTSC	19.64 66	9.36 72	31.06 76	10.99 58	6.24 51	28.10 64	10.14 41	37.50 72	58.34 101	31.74 66	36.70 49	32.50 47	21.03 49	36.00 86	37.24 91	1.23 59	0.44 51	0.12 14	0.31 37	1.98 48	1.69 49

NCC-stereo	19.75 67	6.08 56	23.96 64	15.05 76	16.17 97	18.42 32	21.45 81	43.31 92	41.00 61	24.30 36	38.12 56	34.98 60	26.92 62	35.46 81	25.65 66	4.29 86	2.07 80	3.24 81	8.90 98	2.86 59	2.86 65

Abc-Net	19.75 67	6.08 56	23.96 64	15.05 76	16.17 97	18.42 32	21.45 81	43.31 92	41.00 61	24.30 36	38.12 56	34.98 60	26.92 62	35.46 81	25.65 66	4.29 86	2.07 80	3.24 81	8.90 98	2.86 59	2.86 65

RGC	19.90 69	13.26 88	20.24 61	18.19 87	14.06 95	21.21 43	14.33 56	34.97 57	43.60 70	27.73 51	41.49 69	39.76 71	27.98 69	34.52 78	21.37 43	3.66 83	0.54 57	8.67 97	4.16 89	4.19 72	4.02 75

DeepPrunerF	19.92 70	11.02 85	44.29 91	20.74 91	17.75 100	19.19 36	22.57 83	36.86 68	49.93 82	25.48 43	36.62 48	24.69 28	23.97 53	23.44 44	21.79 45	2.68 78	1.63 75	5.70 89	3.95 88	3.38 66	2.67 63

WCMA_ROB	20.02 71	4.30 39	19.72 59	9.47 46	7.00 60	32.71 79	13.99 54	31.97 39	32.48 37	41.51 87	52.00 93	44.09 84	36.14 88	32.43 71	24.29 58	6.19 95	2.79 90	1.09 54	1.10 62	3.95 69	3.13 68

SANet	20.96 72	6.60 62	29.81 72	8.83 40	3.19 20	31.27 73	20.56 79	41.86 84	56.09 97	39.30 83	43.62 79	44.95 86	31.93 80	27.83 57	23.67 51	0.94 49	0.52 56	0.99 51	0.42 44	4.72 74	2.03 56

FBW_ROB	21.00 73	10.23 79	22.72 62	13.71 68	6.77 57	30.49 72	16.46 65	44.58 94	43.57 69	44.25 93	39.59 61	43.27 83	26.12 60	33.96 76	21.80 46	2.58 77	1.84 78	7.14 96	2.78 81	4.00 70	4.06 76

G-Net	21.25 74	8.24 69	37.97 84	15.14 78	9.26 73	50.87 97	16.80 67	27.96 20	30.54 29	38.92 81	42.93 76	33.27 51	32.17 81	27.99 59	25.25 63	4.40 89	3.83 96	3.60 84	0.97 59	7.91 87	6.95 90

STTStereo_v2	21.25 74	8.24 69	37.97 84	15.14 78	9.26 73	50.87 97	16.80 67	27.96 20	30.54 29	38.92 81	42.93 76	33.27 51	32.17 81	27.99 59	25.25 63	4.40 89	3.83 96	3.60 84	0.97 59	7.91 87	6.95 90

SHD	21.32 76	10.03 78	30.16 74	14.30 72	8.68 69	24.15 54	8.93 34	40.82 81	61.17 103	35.79 77	44.15 81	38.73 68	30.20 75	33.88 74	31.09 82	2.00 67	0.81 65	1.17 57	1.55 69	4.08 71	4.78 82

CSAN	21.34 77	9.45 74	23.34 63	20.99 93	4.95 39	32.57 77	34.26 100	38.83 77	49.95 84	36.97 79	39.72 62	44.97 87	31.73 79	26.05 54	23.94 53	1.52 62	0.47 53	0.85 49	1.43 67	2.73 57	2.14 60

ADCL	21.64 78	6.20 58	47.32 94	9.93 48	6.91 59	38.69 91	19.97 77	31.26 34	54.04 91	27.89 52	47.97 88	31.35 44	30.66 77	33.18 73	36.62 88	0.99 51	0.92 70	2.63 75	1.82 74	2.45 55	2.13 58

ADCPNet	21.93 79	9.38 73	57.92 102	11.76 63	6.88 58	36.03 85	18.44 72	32.80 45	35.93 46	32.91 72	43.80 80	39.21 69	26.99 64	31.14 67	36.94 90	2.05 68	2.67 89	2.30 72	3.45 85	4.22 73	3.82 73

MeshStereo	22.27 80	6.87 64	11.15 27	7.69 31	4.87 37	37.70 88	13.06 53	41.64 83	36.95 50	50.92 98	53.41 94	58.12 99	41.93 95	37.73 89	27.62 75	2.52 76	2.37 86	2.99 79	2.07 78	3.09 62	2.60 62
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
ADCMid	22.76 81	10.75 81	41.73 89	11.23 62	7.97 67	27.26 62	19.18 76	35.44 61	38.02 55	40.23 84	46.38 86	36.56 63	42.70 96	38.20 90	40.75 96	1.77 64	1.43 73	3.05 80	3.81 87	4.96 76	3.82 73

MSMD_ROB	22.90 82	9.64 77	14.76 41	17.46 85	9.88 75	34.94 82	18.83 73	33.67 48	37.47 51	40.97 86	59.80 99	45.93 88	38.34 91	31.74 70	23.24 50	6.22 96	3.58 95	8.78 98	11.03 100	6.76 84	5.06 84

LE_ROB	22.93 83	3.07 27	14.02 34	7.70 32	2.86 16	31.99 76	17.35 70	28.10 23	67.19 110	70.51 110	55.61 98	49.07 93	47.62 100	24.76 50	36.49 87	0.35 23	0.20 31	0.27 23	0.55 49	0.48 10	0.44 19

AnyNet_C32	23.37 84	13.28 89	40.29 88	12.36 64	11.44 84	30.39 69	29.15 96	31.36 36	44.90 74	35.36 75	48.86 89	33.61 53	34.20 84	43.15 95	41.62 99	1.27 60	1.35 72	1.18 58	2.58 80	4.77 75	6.27 86

SGM-ForestM	23.72 85	2.44 18	9.13 18	7.43 26	2.25 10	44.80 93	19.11 75	44.90 95	49.97 85	50.74 97	51.18 92	62.06 106	45.83 99	44.96 96	33.88 84	1.00 52	0.84 66	0.78 45	0.62 51	1.29 36	1.11 33

MFN_U_SF_DS_RVC	23.92 86	13.60 90	30.09 73	25.46 99	10.02 77	38.51 90	20.40 78	35.77 62	43.76 71	42.15 88	47.10 87	39.99 73	28.20 71	35.96 85	27.06 72	3.64 82	3.85 98	6.66 92	16.30 104	4.97 77	4.98 83

XQC	24.10 87	16.76 92	50.68 97	21.37 95	11.01 81	35.24 83	18.84 74	37.28 71	55.11 94	31.64 65	30.06 23	37.71 65	30.31 76	37.17 87	39.66 94	4.20 85	0.41 46	2.76 78	1.89 75	11.46 98	8.42 96

FC-DCNN	24.37 88	10.99 84	19.02 58	18.44 88	9.16 72	36.98 86	23.07 86	40.44 80	43.05 66	46.25 96	53.58 95	50.44 94	37.86 89	35.45 80	27.57 74	6.87 97	3.37 92	6.01 90	4.88 92	7.86 86	6.05 85

DPSNet	24.40 89	6.97 66	33.14 79	11.16 61	6.54 54	53.33 100	43.32 106	51.28 102	59.37 102	30.89 61	39.36 59	40.35 75	34.30 85	32.57 72	25.09 62	4.65 92	1.62 74	0.37 28	0.66 52	8.51 91	4.44 80

PDISCO_ROB	24.45 90	9.24 71	29.28 71	28.68 102	19.96 101	37.14 87	15.33 61	45.04 96	54.69 93	29.24 56	42.67 73	46.07 89	28.17 70	35.35 79	30.30 79	9.92 101	2.13 82	6.90 93	3.20 83	8.74 92	6.87 89

GANetREF_RVC	25.41 91	29.28 103	24.63 66	25.18 98	6.04 50	30.44 70	35.39 102	47.84 97	50.60 88	36.88 78	36.04 47	34.11 56	31.23 78	37.46 88	27.96 78	8.58 98	2.66 88	16.29 103	4.95 93	14.37 102	8.32 95
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
ADCS	26.20 92	13.71 91	46.90 93	14.51 74	10.34 79	30.46 71	22.72 85	42.81 90	57.61 100	42.58 89	49.37 90	41.91 80	40.43 94	41.90 94	44.58 103	2.37 72	2.19 84	2.37 73	3.29 84	7.32 85	6.52 88

PASM	28.19 93	17.35 93	48.18 95	22.49 96	22.20 102	25.92 59	25.84 92	34.35 52	49.45 80	35.37 76	39.48 60	46.72 91	33.36 83	34.28 77	36.77 89	11.73 104	14.94 105	17.43 105	22.18 106	13.45 99	12.32 102

AnyNet_C01	29.76 94	25.78 102	75.37 108	17.03 84	11.34 83	49.00 96	27.12 94	37.99 74	40.05 60	40.73 85	65.13 108	46.98 92	38.86 93	46.63 97	48.19 105	2.14 69	2.13 82	2.13 67	2.56 79	8.14 90	7.98 94

LSM	29.98 95	19.52 98	55.11 100	19.06 89	53.75 109	32.57 77	23.17 87	42.61 87	48.57 78	45.97 95	55.04 97	42.93 82	36.02 87	35.63 83	27.03 71	2.33 70	9.19 104	4.87 88	7.71 97	10.65 96	27.91 107

RTSA	30.15 96	23.12 100	96.33 111	16.71 82	11.63 85	58.66 104	24.28 90	36.80 66	62.63 105	43.79 91	46.10 84	32.56 48	43.90 97	51.29 101	40.92 97	2.38 73	0.66 60	1.06 52	0.74 54	5.11 78	4.34 78

RTS	30.15 96	23.12 100	96.33 111	16.71 82	11.63 85	58.66 104	24.28 90	36.80 66	62.63 105	43.79 91	46.10 84	32.56 48	43.90 97	51.29 101	40.92 97	2.38 73	0.66 60	1.06 52	0.74 54	5.11 78	4.34 78

DispFullNet	30.27 98	31.74 104	58.63 103	30.13 103	28.18 104	31.79 75	8.94 35	37.16 70	54.68 92	32.60 71	37.99 53	44.62 85	28.80 72	41.55 93	27.93 77	11.94 105	3.46 93	35.53 107	12.34 101	30.08 106	17.26 103

MANE	31.97 99	10.75 81	17.71 55	20.94 92	13.77 93	47.80 95	26.85 93	53.00 106	55.20 95	59.29 101	64.30 106	64.93 108	56.31 107	51.83 104	38.69 93	9.61 100	4.72 99	6.91 94	19.93 105	9.20 93	7.64 93

PVD	33.19 100	18.59 95	37.64 83	26.09 100	16.99 99	33.25 80	42.25 105	51.42 103	66.03 108	52.65 100	64.31 107	51.93 97	55.13 104	51.61 103	40.40 95	4.00 84	3.51 94	9.53 99	4.26 90	10.62 95	23.56 104

ELAS	33.68 101	18.95 96	36.71 81	20.64 90	13.61 92	56.43 101	35.14 101	50.03 101	49.90 81	60.58 103	63.83 105	58.64 102	53.36 101	50.73 99	44.44 102	11.03 102	6.58 102	10.02 100	7.43 96	13.58 101	11.94 101
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ELAS_RVC	33.79 102	19.15 97	35.97 80	21.28 94	13.79 94	52.58 99	36.11 103	48.96 100	55.66 96	61.47 106	62.91 101	58.35 101	53.55 102	53.37 106	42.23 100	11.25 103	6.64 103	10.32 101	7.05 95	13.56 100	11.58 100
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SAMSARA	34.43 103	18.53 94	31.91 77	55.34 109	35.34 106	75.99 110	94.71 113	47.97 98	50.84 89	37.47 80	36.99 50	53.12 98	38.55 92	39.56 92	38.65 92	3.59 81	5.40 101	2.22 71	3.72 86	9.43 94	9.29 99

BEATNet-Init1	35.08 104	20.64 99	54.70 99	27.43 101	12.00 89	58.29 103	31.01 97	52.58 105	57.49 99	60.36 102	62.98 102	69.98 110	55.54 105	53.00 105	44.40 101	4.61 91	2.84 91	7.05 95	6.89 94	10.97 97	8.94 97

MSC_U_SF_DS_RVC	37.16 105	34.69 106	53.68 98	37.47 105	12.55 91	56.80 102	22.63 84	52.02 104	63.29 107	66.20 108	63.82 104	58.12 99	53.73 103	51.21 100	46.48 104	8.79 99	5.31 100	14.69 102	15.22 103	17.27 104	9.21 98

NVStereoNet_ROB	41.93 106	32.40 105	44.55 92	36.60 104	27.78 103	34.52 81	32.38 98	48.84 99	57.18 98	69.34 109	75.62 112	66.07 109	64.86 111	47.34 98	60.78 108	15.72 106	23.99 106	23.40 106	32.06 107	21.34 105	23.86 105
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
MADNet+	53.31 107	68.61 109	98.75 113	67.03 110	37.58 107	79.83 112	81.74 111	60.38 109	61.61 104	51.43 99	54.92 96	61.11 105	59.29 109	69.66 110	76.71 111	23.88 107	24.03 107	17.08 104	12.35 102	32.47 107	27.69 106

SGM+DAISY	54.67 108	56.61 107	62.73 105	40.21 106	53.22 108	61.95 107	48.59 108	55.05 107	44.62 73	61.03 105	63.68 103	60.46 104	55.62 106	60.68 108	55.34 106	56.34 108	48.56 110	50.16 108	49.63 110	52.25 108	56.76 110

SPS-STEREO	55.62 109	59.14 108	64.16 106	45.05 107	53.84 110	59.88 106	44.00 107	59.53 108	49.94 83	62.33 107	61.09 100	59.80 103	56.82 108	60.06 107	57.85 107	58.41 109	48.34 109	51.07 109	49.21 109	55.41 109	56.48 109
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
PWCK	65.90 110	88.63 112	79.40 109	80.32 112	29.96 105	63.85 108	39.45 104	72.07 111	71.26 111	78.74 111	68.99 109	77.82 111	61.77 110	82.81 112	62.64 110	80.17 110	39.88 108	81.90 112	47.37 108	73.54 111	37.51 108

MFMNet_re	72.05 111	79.64 110	64.23 107	52.73 108	84.71 112	77.93 111	68.35 109	63.96 110	66.93 109	61.00 104	71.29 110	63.84 107	68.91 112	72.42 111	61.11 109	82.03 112	74.22 112	80.34 111	74.44 111	83.62 112	89.34 112

DPSimNet_ROB	74.29 112	83.27 111	59.30 104	77.19 111	65.13 111	75.14 109	74.95 110	73.99 112	76.94 112	81.99 112	75.48 111	80.18 112	71.76 113	67.39 109	84.00 112	81.87 111	51.14 111	79.99 110	76.70 112	70.17 110	79.21 111

MADNet++	94.16 113	94.02 113	90.25 110	96.28 113	96.84 113	97.17 113	90.16 112	94.71 113	91.69 113	97.07 113	93.71 113	94.62 113	92.63 114	96.15 113	95.18 113	95.42 113	95.89 113	92.80 113	92.23 113	92.10 113	94.30 113

MEDIAN_ROB	99.19 114	99.84 114	99.62 115	98.49 114	98.51 114	98.58 114	97.81 115	98.80 114	98.56 114	99.36 116	99.49 114	99.56 114	99.06 115	98.35 114	98.31 114	99.99 114	99.63 114	100.00 114	100.00 114	99.81 114	99.95 114

AVERAGE_ROB	99.80 115	99.99 115	99.40 114	100.00 115	100.00 115	98.99 115	97.72 114	100.00 115	100.00 115	100.00 117	100.00 115	100.00 117	100.00 116	100.00 117	100.00 117	100.00 115	100.00 115	100.00 114	100.00 114	100.00 115	100.00 115

DPSMNet_ROB	99.95 116	100.00 116	100.00 116	100.00 115	100.00 115	100.00 116	100.00 116	100.00 115	100.00 115	99.15 115	100.00 115	99.96 115	100.00 116	99.99 115	99.99 115	100.00 115	100.00 115	100.00 114	100.00 114	100.00 115	100.00 115

DGTPSM_ROB	99.95 116	100.00 116	100.00 116	100.00 115	100.00 115	100.00 116	100.00 116	100.00 115	100.00 115	99.14 114	100.00 115	99.96 115	100.00 116	99.99 115	99.99 115	100.00 115	100.00 115	100.00 114	100.00 114	100.00 115	100.00 115

DPSM_ROB	100.00 118	100.00 116	100.00 116	100.00 115	100.00 115	100.00 116	100.00 116	100.00 115	100.00 115	100.00 117	100.00 115	100.00 117	100.00 116	100.00 117	100.00 117	100.00 115	100.00 115	100.00 114	100.00 114	100.00 115	100.00 115

LSM0	100.00 118	100.00 116	100.00 116	100.00 115	100.00 115	100.00 116	100.00 116	100.00 115	100.00 115	100.00 117	100.00 115	100.00 117	100.00 116	100.00 117	100.00 117	100.00 115	100.00 115	100.00 114	100.00 114	100.00 115	100.00 115

DPSM	100.00 118	100.00 116	100.00 116	100.00 115	100.00 115	100.00 116	100.00 116	100.00 115	100.00 115	100.00 117	100.00 115	100.00 117	100.00 116	100.00 117	100.00 117	100.00 115	100.00 115	100.00 114	100.00 114	100.00 115	100.00 115

MSMDNet													7.19 5

Low-res two-view benchmark