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):

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 Infoalllakes. 1llakes. 1ssand box 1lsand box 1sstora. room 1lstora. room 1sstora. room 2lstora. room 2sstora. room 2 1lstora. room 2 1sstora. room 2 2lstora. room 2 2sstora. room 3lstora. room 3stunnel 1ltunnel 1stunnel 2ltunnel 2stunnel 3ltunnel 3s
sort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysorted bysort bysort by
CPSMnettwo views3.41
3
0.64
9
4.57
13
3.35
41
2.45
36
5.61
13
0.36
6
12.09
16
6.09
3
4.58
2
8.66
8
6.37
5
2.23
6
8.24
12
2.63
6
0.02
15
0.02
18
0.00
1
0.00
1
0.10
12
0.09
10
iResNetv2_ROBtwo views4.28
14
1.43
32
7.17
33
2.91
33
1.26
23
4.36
6
1.62
15
13.64
27
10.25
22
9.83
21
11.41
19
7.68
12
4.00
15
7.75
2
1.85
2
0.00
1
0.00
1
0.00
1
0.00
1
0.37
40
0.09
10
DLCB_ROBtwo views4.51
16
0.91
15
3.78
9
2.19
24
1.07
20
6.28
15
3.09
18
9.78
5
7.72
7
10.65
23
12.97
25
13.91
28
3.71
13
8.72
17
5.30
18
0.00
1
0.00
1
0.00
1
0.00
1
0.03
5
0.10
16
CBMV_ROBtwo views4.14
11
0.52
4
3.14
3
1.30
4
0.77
12
6.92
19
1.97
17
10.11
7
9.58
16
8.92
17
14.20
31
7.12
11
5.90
27
8.65
16
3.50
13
0.01
10
0.05
28
0.00
1
0.00
1
0.04
6
0.09
10
MDST_ROBtwo views8.37
37
0.32
1
9.03
40
4.18
46
2.42
35
26.86
52
6.14
29
19.36
44
13.52
28
27.09
47
22.75
38
9.47
19
4.74
19
15.06
42
6.34
20
0.02
15
0.02
18
0.00
1
0.00
1
0.02
1
0.13
19
iResNet_ROBtwo views4.23
13
1.02
19
4.90
15
2.18
23
0.93
17
2.92
1
0.37
7
15.10
33
16.91
39
7.89
16
10.51
15
7.03
8
3.07
11
8.16
11
3.46
12
0.01
10
0.00
1
0.00
1
0.00
1
0.10
12
0.02
1
NOSS_ROBtwo views3.30
2
0.46
3
2.62
1
2.08
21
1.01
19
5.60
12
0.74
9
10.37
8
11.48
24
5.15
4
8.43
7
5.67
4
1.73
3
7.97
5
2.34
5
0.02
15
0.06
32
0.00
1
0.00
1
0.07
9
0.14
20
NCCL2two views5.88
26
1.59
35
5.44
19
1.87
15
0.92
16
9.55
25
11.55
50
12.11
17
9.94
20
9.67
20
8.85
10
22.28
41
7.41
33
8.78
18
7.17
24
0.01
10
0.00
1
0.03
34
0.00
1
0.13
20
0.23
28
DISCOtwo views6.28
30
0.57
5
5.78
23
3.43
42
1.17
22
11.22
34
3.39
21
12.14
18
16.16
37
6.52
9
11.22
18
16.96
31
6.32
28
19.51
47
10.74
40
0.00
1
0.00
1
0.00
1
0.00
1
0.35
38
0.11
17
iResNettwo views3.68
8
0.91
15
7.94
38
2.97
35
0.34
3
4.44
9
0.48
8
7.70
3
9.74
18
7.72
14
12.74
24
4.03
2
2.87
10
8.05
7
3.37
11
0.02
15
0.01
15
0.00
1
0.00
1
0.10
12
0.09
10
DN-CSS_ROBtwo views2.69
1
1.40
31
5.34
18
2.31
26
0.75
10
3.14
2
0.06
1
6.11
1
3.87
1
5.34
6
12.18
21
2.34
1
1.22
1
7.84
3
1.48
1
0.03
24
0.00
1
0.00
1
0.00
1
0.35
38
0.03
6
HSMtwo views4.00
10
0.79
10
3.16
4
1.59
9
2.17
34
6.77
18
1.11
11
12.28
20
6.35
4
6.75
11
8.11
6
13.90
27
5.37
24
8.85
19
2.71
8
0.00
1
0.00
1
0.00
1
0.00
1
0.02
1
0.02
1
MSMD_ROBtwo views9.28
40
1.09
22
4.65
14
1.58
8
0.39
4
16.52
43
4.41
25
13.60
26
14.87
32
22.34
42
39.89
51
25.67
43
20.71
46
12.42
32
6.98
23
0.34
41
0.03
23
0.00
1
0.00
1
0.05
7
0.09
10
MLCVtwo views3.44
5
0.88
14
5.60
20
1.39
6
0.25
1
4.36
6
0.33
4
7.25
2
7.28
6
9.17
18
12.24
22
5.09
3
2.47
9
9.15
23
3.23
10
0.00
1
0.00
1
0.00
1
0.00
1
0.10
12
0.02
1
PWC_ROBbinarytwo views8.24
36
3.13
47
12.74
47
2.43
28
4.43
45
7.51
20
1.22
12
16.63
39
19.24
42
16.08
38
28.29
42
13.99
29
10.16
38
13.63
38
14.06
48
0.42
43
0.00
1
0.05
38
0.00
1
0.59
42
0.27
34
PWCDC_ROBbinarytwo views7.92
35
3.17
49
7.48
37
5.73
49
4.40
44
10.45
32
0.35
5
14.52
31
28.19
48
10.36
22
31.27
45
7.04
9
9.14
35
13.22
35
8.78
33
2.74
51
0.02
18
0.00
1
0.00
1
1.31
47
0.17
24
CC-Nettwo views4.98
20
1.47
33
7.42
36
2.40
27
2.14
33
8.73
22
3.64
22
12.42
21
13.11
27
7.03
12
7.57
5
7.88
15
6.52
30
10.16
27
7.82
25
0.02
15
0.03
23
0.00
1
0.00
1
0.11
17
1.07
45
ccs_robtwo views3.57
7
1.21
26
3.53
6
1.85
13
1.09
21
4.72
10
0.23
3
11.67
14
8.23
10
5.21
5
9.81
13
6.65
6
2.44
8
9.14
22
5.54
19
0.00
1
0.00
1
0.00
1
0.00
1
0.16
21
0.02
1
ccstwo views3.68
8
1.69
38
5.15
16
2.56
30
2.78
39
6.23
14
1.83
16
11.76
15
8.34
12
2.73
1
5.74
1
8.98
17
3.51
12
8.61
15
2.16
3
0.69
45
0.36
44
0.21
44
0.00
1
0.31
34
0.04
7
MeshStereopermissivetwo views11.52
45
1.52
34
4.55
12
1.89
16
1.46
25
19.87
47
5.11
27
20.66
48
15.91
36
32.67
51
34.51
49
39.34
54
21.15
47
18.74
46
12.10
44
0.11
35
0.06
32
0.01
30
0.00
1
0.45
41
0.22
27
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
AdaStereotwo views3.41
3
0.58
6
5.33
17
3.52
43
0.85
14
4.08
5
1.29
14
12.16
19
7.77
8
6.57
10
9.62
12
7.82
13
1.53
2
4.85
1
2.20
4
0.00
1
0.02
18
0.00
1
0.00
1
0.02
1
0.02
1
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
CBMVpermissivetwo views5.35
22
0.91
15
3.67
8
1.62
10
0.44
6
10.09
29
7.19
39
12.49
22
12.33
26
12.22
27
14.69
32
10.93
21
6.48
29
8.51
13
4.96
17
0.02
15
0.15
41
0.00
1
0.00
1
0.17
23
0.17
24
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
CPSMnet_ROBtwo views4.83
18
0.85
12
7.04
32
3.28
39
4.53
48
11.41
35
8.66
41
16.16
37
9.34
15
5.34
6
5.75
2
8.15
16
2.43
7
9.05
21
4.38
15
0.00
1
0.00
1
0.00
1
0.00
1
0.09
11
0.06
8
pmcnntwo views7.72
33
1.27
29
9.42
42
2.91
33
3.14
40
9.44
24
6.23
30
12.56
24
16.51
38
14.53
36
24.08
39
27.44
46
8.49
34
9.32
24
8.44
30
0.06
30
0.08
37
0.00
1
0.00
1
0.30
33
0.15
21
SGM-Foresttwo views4.96
19
0.32
1
2.84
2
1.21
2
0.64
7
10.23
31
6.64
35
11.55
12
10.98
23
10.94
24
13.59
27
11.65
24
4.30
17
8.94
20
4.63
16
0.11
35
0.04
26
0.00
1
0.00
1
0.05
7
0.46
38
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
WCMA_ROBtwo views9.21
39
0.87
13
7.37
35
2.54
29
2.13
32
13.59
40
5.80
28
11.64
13
14.01
29
24.43
45
32.99
47
27.09
45
18.02
41
12.51
33
9.85
38
0.81
47
0.07
35
0.01
30
0.01
26
0.16
21
0.23
28
DeepPruner_ROBtwo views3.52
6
1.14
23
4.06
11
1.12
1
1.65
27
3.65
3
0.83
10
13.96
28
4.47
2
7.80
15
10.84
16
7.05
10
2.16
5
8.14
10
3.08
9
0.07
31
0.03
23
0.00
1
0.01
26
0.32
35
0.06
8
SUWNettwo views4.16
12
1.16
24
3.60
7
1.94
20
1.46
25
3.92
4
4.68
26
10.89
11
8.34
12
7.58
13
10.84
16
10.27
20
6.62
31
8.56
14
2.69
7
0.39
42
0.00
1
0.00
1
0.01
26
0.21
29
0.09
10
LALA_ROBtwo views6.58
31
1.80
41
6.25
26
1.26
3
0.94
18
10.08
28
9.02
42
16.00
36
11.51
25
12.74
32
13.02
26
24.77
42
5.25
22
10.56
29
8.02
26
0.04
25
0.05
28
0.00
1
0.02
29
0.10
12
0.25
31
SANettwo views10.64
44
1.86
42
10.91
44
1.76
11
0.71
9
14.62
42
9.23
44
19.18
42
37.14
55
19.22
40
27.96
41
25.86
44
19.11
45
13.02
34
10.63
39
0.08
32
0.06
32
0.03
34
0.02
29
0.62
43
0.81
44
ETE_ROBtwo views5.80
24
1.77
40
6.33
28
1.44
7
0.78
13
6.43
17
6.90
36
12.53
23
8.08
9
12.93
34
14.89
33
21.13
40
5.87
26
9.83
25
6.57
22
0.04
25
0.01
15
0.00
1
0.02
29
0.08
10
0.33
36
StereoDRNettwo views5.59
23
1.75
39
6.80
30
3.12
37
4.45
46
10.61
33
4.35
24
18.80
40
9.73
17
12.22
27
6.87
3
11.44
23
4.65
18
8.09
9
8.26
28
0.02
15
0.11
39
0.00
1
0.03
32
0.20
26
0.28
35
XPNet_ROBtwo views6.03
28
1.22
27
5.61
21
2.56
30
0.90
15
6.32
16
7.07
37
12.92
25
8.30
11
14.76
37
15.13
35
19.84
37
6.66
32
10.36
28
8.58
31
0.02
15
0.04
26
0.00
1
0.03
32
0.11
17
0.24
30
CSANtwo views7.62
32
1.60
36
6.56
29
1.83
12
0.66
8
12.40
37
10.52
49
14.45
30
21.32
43
14.19
35
15.98
36
17.84
32
13.02
40
12.32
31
8.38
29
0.09
33
0.07
35
0.03
34
0.04
34
0.33
36
0.67
42
DRN-Testtwo views5.87
25
0.98
18
5.89
24
2.69
32
3.65
43
12.37
36
3.35
20
20.07
47
10.20
21
11.93
26
12.31
23
11.06
22
5.31
23
7.89
4
9.05
34
0.04
25
0.05
28
0.04
37
0.04
34
0.18
24
0.25
31
StereoDRNet-Refinedtwo views4.46
15
0.62
8
3.80
10
1.92
18
0.40
5
9.35
23
0.15
2
10.02
6
8.83
14
12.69
31
11.62
20
9.34
18
3.87
14
8.06
8
8.02
26
0.00
1
0.00
1
0.01
30
0.05
36
0.20
26
0.26
33
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
Anonymous Stereotwo views6.16
29
3.15
48
23.75
54
2.97
35
2.48
37
4.39
8
13.30
52
9.21
4
9.86
19
9.56
19
8.76
9
6.79
7
1.99
4
13.50
36
13.04
47
0.01
10
0.05
28
0.00
1
0.06
37
0.22
30
0.19
26
NVstereo2Dtwo views4.51
16
0.82
11
6.86
31
3.28
39
3.38
42
8.16
21
3.13
19
10.51
9
15.15
33
4.90
3
6.89
4
7.87
14
4.78
20
9.88
26
3.91
14
0.01
10
0.00
1
0.00
1
0.06
37
0.02
1
0.58
40
FC-DCNNcopylefttwo views16.43
51
1.17
25
7.26
34
3.66
44
2.52
38
28.02
53
8.49
40
27.36
54
28.40
49
42.84
55
48.34
57
48.13
56
35.36
56
26.27
53
20.04
51
0.10
34
0.02
18
0.24
46
0.07
39
0.19
25
0.15
21
PDISCO_ROBtwo views9.62
41
1.99
44
11.51
45
9.88
54
9.61
53
21.48
48
3.83
23
19.33
43
28.49
50
11.27
25
14.17
30
19.92
38
5.02
21
16.35
44
9.18
35
5.28
53
0.41
46
0.14
43
0.09
40
2.05
51
2.36
50
PSMNet_ROBtwo views5.02
21
1.63
37
6.03
25
1.90
17
1.83
31
9.57
26
6.35
33
15.58
35
7.23
5
6.15
8
10.48
14
12.22
26
4.16
16
8.02
6
8.71
32
0.02
15
0.01
15
0.01
30
0.10
41
0.20
26
0.12
18
NaN_ROBtwo views6.00
27
1.24
28
6.29
27
1.34
5
1.68
29
9.60
27
10.31
47
15.09
32
15.79
35
12.62
30
8.95
11
11.67
25
5.83
25
11.78
30
6.41
21
0.05
28
0.13
40
0.08
39
0.20
42
0.22
30
0.79
43
LE_ROBtwo views16.73
54
1.28
30
11.61
46
3.72
45
1.65
27
16.67
44
9.17
43
14.39
29
55.91
57
63.81
57
40.86
54
35.94
51
37.73
57
14.24
40
26.87
56
0.05
28
0.10
38
0.13
42
0.22
43
0.12
19
0.15
21
MFMNet_retwo views13.29
46
8.60
55
18.29
51
9.75
53
7.25
51
19.65
46
14.84
53
20.71
49
30.72
52
23.03
43
28.77
43
18.85
35
26.09
52
13.55
37
9.82
36
2.44
50
1.35
52
0.34
49
0.23
44
4.78
53
6.69
52
SGM_ROBbinarytwo views10.08
42
0.60
7
3.42
5
2.30
25
0.32
2
19.41
45
6.33
32
18.95
41
14.64
30
25.14
46
24.32
40
33.34
49
18.79
44
19.86
48
12.55
46
0.25
39
0.26
43
0.22
45
0.24
45
0.34
37
0.40
37
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DPSNettwo views10.14
43
1.88
43
16.82
50
1.85
13
1.73
30
24.84
50
17.20
56
19.92
46
27.41
47
12.23
29
13.62
28
16.52
30
18.35
42
14.42
41
12.50
45
0.78
46
0.54
49
0.08
39
0.25
46
1.18
46
0.59
41
LSMtwo views14.01
47
5.95
51
33.49
55
6.78
51
43.61
57
10.22
30
9.98
46
15.16
34
22.93
44
23.07
44
32.34
46
18.52
34
12.67
39
15.45
43
11.10
43
0.16
37
0.51
48
0.09
41
0.32
47
1.08
45
16.85
57
SGM-ForestMtwo views16.99
55
1.08
21
5.74
22
2.12
22
0.75
10
31.63
55
12.21
51
27.80
56
32.25
53
37.88
52
39.99
52
52.96
57
35.20
55
33.60
56
24.47
55
0.26
40
0.39
45
0.31
48
0.39
48
0.26
32
0.53
39
ELAS_ROBcopylefttwo views16.54
52
2.26
46
10.09
43
5.50
48
4.46
47
28.28
54
16.72
55
25.55
53
33.54
54
40.19
53
40.30
53
36.68
52
30.03
53
29.40
55
20.61
52
0.98
49
1.21
50
0.86
51
0.70
49
1.39
48
2.16
48
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
FBW_ROBtwo views8.50
38
1.03
20
7.98
39
1.93
19
1.28
24
13.10
39
6.23
30
22.50
50
18.98
40
18.82
39
14.91
34
19.06
36
10.04
37
18.41
45
9.83
37
0.62
44
0.22
42
1.82
52
0.82
50
0.99
44
1.36
47
ELAScopylefttwo views16.72
53
2.14
45
9.23
41
4.92
47
4.53
48
32.66
56
15.11
54
27.40
55
28.68
51
40.27
54
44.90
55
38.33
53
30.50
54
26.44
54
21.94
54
0.88
48
1.23
51
0.67
50
0.89
51
1.49
49
2.18
49
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
PASMtwo views7.90
34
4.22
50
21.97
53
3.25
38
3.29
41
5.39
11
6.57
34
10.57
10
19.09
41
12.77
33
13.92
29
18.11
33
9.51
36
13.79
39
10.77
42
0.19
38
0.45
47
0.29
47
1.08
52
1.49
49
1.19
46
SPS-STEREOcopylefttwo views15.04
48
6.23
52
13.21
49
11.34
55
11.65
55
23.30
49
7.15
38
24.16
52
15.65
34
31.78
50
29.19
44
31.62
47
21.32
48
24.62
51
19.50
50
7.59
55
4.19
56
3.22
53
1.48
53
6.99
55
6.54
51
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
SGM+DAISYtwo views15.62
49
7.26
54
19.28
52
8.94
52
10.11
54
26.25
51
10.49
48
19.36
44
14.65
31
30.64
49
33.59
48
33.00
48
22.32
49
24.96
52
16.42
49
7.90
56
6.25
57
4.51
54
3.37
54
5.86
54
7.20
53
PWCKtwo views30.53
57
44.32
57
47.25
58
29.76
57
7.23
50
40.78
57
27.10
57
44.73
57
44.32
56
47.31
56
36.37
50
47.16
55
26.05
51
41.26
57
31.87
57
21.83
57
4.03
55
29.50
56
4.67
55
27.17
57
7.80
54
DispFullNettwo views17.47
56
26.01
56
33.98
56
22.58
56
20.86
56
13.84
41
1.28
13
16.50
38
26.27
46
19.97
41
17.17
37
20.52
39
18.49
43
22.86
50
10.76
41
5.13
52
2.83
53
30.72
57
7.72
56
20.86
56
11.01
56
NVStereoNet_ROBtwo views16.04
50
6.75
53
12.90
48
6.37
50
7.42
52
12.89
38
9.74
45
22.78
51
25.12
45
30.32
48
46.19
56
34.37
50
25.38
50
21.48
49
21.38
53
5.94
54
3.10
54
6.07
55
10.09
57
4.01
52
8.54
55
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
DPSimNet_ROBtwo views53.45
58
64.73
58
44.39
57
53.97
58
45.39
58
53.66
58
54.83
58
55.15
58
57.87
58
64.16
58
50.83
58
63.40
58
53.34
58
46.45
58
65.81
58
63.13
58
26.54
58
57.94
58
51.11
58
45.52
58
50.69
58
MEDIAN_ROBtwo views98.41
59
99.70
59
99.30
60
97.09
59
97.02
59
96.89
59
95.77
60
97.66
59
97.28
59
98.79
61
98.94
59
99.18
59
98.14
59
96.89
59
96.88
59
99.96
61
99.16
59
100.00
59
99.99
59
99.69
59
99.88
59
AVERAGE_ROBtwo views99.62
60
99.95
60
98.81
59
100.00
64
100.00
60
98.08
60
95.47
59
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
99.99
60
100.00
63
100.00
60
100.00
59
100.00
60
100.00
62
100.00
63
DPSMtwo views99.95
63
100.00
61
100.00
63
99.76
60
100.00
60
100.00
61
100.00
61
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
63
99.21
59
100.00
60
100.00
59
100.00
60
99.99
60
99.95
60
DPSM_ROBtwo views99.95
63
100.00
61
100.00
63
99.76
60
100.00
60
100.00
61
100.00
61
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
63
99.21
59
100.00
60
100.00
59
100.00
60
99.99
60
99.95
60
DGTPSM_ROBtwo views99.90
61
100.00
61
99.99
61
99.99
62
100.00
60
100.00
61
100.00
61
99.97
60
100.00
60
98.35
59
100.00
60
99.84
60
100.00
60
99.98
60
99.99
60
99.99
62
100.00
60
100.00
59
100.00
60
100.00
62
100.00
63
DPSMNet_ROBtwo views99.91
62
100.00
61
99.99
61
99.99
62
100.00
60
100.00
61
100.00
61
99.98
61
100.00
60
98.35
59
100.00
60
99.84
60
100.00
60
99.98
60
99.99
60
100.00
63
100.00
60
100.00
59
100.00
60
100.00
62
100.00
63
LSM0two views100.00
65
100.00
61
100.00
63
100.00
64
100.00
60
100.00
61
100.00
61
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
60
100.00
62
100.00
63
100.00
63
100.00
60
100.00
59
100.00
60
100.00
62
99.99
62