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
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
R-Stereo Traintwo views7.04
1
1.34
2
7.89
10
6.23
13
2.30
7
7.51
2
5.54
12
23.96
7
6.94
1
16.26
5
33.57
29
12.04
2
5.61
1
7.67
1
3.74
1
0.01
1
0.05
10
0.01
4
0.00
1
0.10
1
0.12
3
R-Stereotwo views7.04
1
1.34
2
7.89
10
6.23
13
2.30
7
7.51
2
5.54
12
23.96
7
6.94
1
16.26
5
33.57
29
12.04
2
5.61
1
7.67
1
3.74
1
0.01
1
0.05
10
0.01
4
0.00
1
0.10
1
0.12
3
HITNettwo views7.83
3
2.93
20
8.39
13
4.76
1
1.48
3
12.75
5
4.64
7
20.97
1
14.37
3
15.14
3
22.15
2
14.57
4
10.62
7
14.38
6
8.14
8
0.04
4
0.01
2
0.48
29
0.02
11
0.57
14
0.10
2
DN-CSS_ROBtwo views8.17
4
2.59
16
14.14
30
6.11
10
3.22
18
6.31
1
3.53
3
22.92
6
20.59
6
16.66
7
30.05
16
9.31
1
7.95
5
13.49
4
4.78
3
0.72
31
0.00
1
0.45
27
0.00
1
0.54
12
0.09
1
MLCVtwo views9.32
5
2.22
12
12.55
25
5.22
3
1.32
1
13.79
7
1.35
1
21.14
2
23.65
9
21.91
18
31.37
20
17.14
6
7.57
4
16.00
13
10.18
13
0.22
14
0.01
2
0.03
8
0.02
11
0.40
6
0.29
11
ccstwo views9.81
6
2.28
13
6.63
4
6.21
12
2.80
12
16.72
21
5.61
15
21.25
3
22.08
8
18.83
11
29.47
14
28.64
27
12.54
10
15.17
11
6.62
5
0.06
7
0.09
22
0.39
24
0.09
21
0.48
7
0.29
11
CFNet_RVCtwo views9.87
7
2.18
11
5.49
2
7.45
22
5.25
35
15.76
15
5.98
16
21.89
4
21.87
7
14.96
1
30.02
15
25.68
23
13.68
13
12.46
3
10.94
18
0.08
8
0.06
16
1.83
54
0.22
26
1.09
25
0.43
15
AdaStereotwo views10.22
8
3.63
29
9.14
18
9.15
32
3.24
19
15.79
16
5.39
11
31.94
28
25.42
10
18.39
9
26.47
5
19.44
8
9.50
6
16.60
16
8.59
9
0.44
25
0.05
10
0.40
25
0.00
1
0.57
14
0.16
7
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
iResNettwo views10.26
9
3.10
25
15.72
38
7.35
20
2.13
5
13.86
8
7.07
23
22.80
5
28.01
16
20.10
13
30.40
18
16.79
5
11.06
8
14.46
7
10.29
16
0.39
23
0.05
10
0.00
1
0.07
20
0.73
18
0.73
20
ccs_robtwo views10.52
10
2.12
9
7.47
9
6.02
9
3.07
15
19.32
26
4.16
4
25.06
9
27.77
15
20.21
14
27.93
12
30.56
32
13.32
11
14.95
9
7.25
6
0.04
4
0.05
10
0.21
14
0.05
16
0.55
13
0.27
10
CFNettwo views10.67
11
2.33
14
8.90
15
6.66
17
4.09
23
16.05
17
4.26
5
30.44
23
31.01
23
18.53
10
26.48
6
23.69
16
13.68
13
15.27
12
10.59
17
0.05
6
0.02
5
0.44
26
0.05
16
0.73
18
0.23
8
DeepPruner_ROBtwo views10.77
12
4.56
38
13.13
26
6.37
15
4.28
26
10.14
4
6.86
21
36.42
49
19.59
5
17.87
8
27.37
10
23.19
12
12.26
9
19.09
22
10.23
15
1.05
43
0.48
45
0.23
16
0.15
24
0.89
23
1.17
30
HSM-Net_RVCpermissivetwo views10.88
13
1.23
1
5.46
1
5.57
4
2.63
10
20.12
27
6.06
18
27.67
14
28.35
17
20.63
15
25.51
3
34.01
41
14.38
19
16.18
14
9.29
12
0.11
9
0.07
17
0.01
4
0.00
1
0.13
3
0.14
6
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
NOSS_ROBtwo views10.99
14
3.81
33
6.88
6
6.69
18
2.30
7
16.06
18
10.94
33
30.06
21
32.41
26
16.10
4
18.72
1
22.78
10
14.21
17
17.52
18
8.62
10
2.35
60
1.84
68
3.32
71
1.57
56
2.06
43
1.52
35
HSMtwo views11.33
15
1.80
5
7.04
8
6.13
11
3.99
22
16.16
20
6.92
22
29.69
19
19.39
4
22.95
22
26.05
4
41.08
61
15.79
23
20.25
26
8.89
11
0.02
3
0.03
7
0.01
4
0.00
1
0.15
4
0.34
14
iResNet_ROBtwo views11.71
16
1.96
8
10.77
20
6.01
8
3.68
20
15.13
14
2.34
2
31.72
27
37.58
40
26.16
32
32.77
26
25.88
24
15.34
21
16.35
15
7.94
7
0.19
12
0.01
2
0.00
1
0.00
1
0.31
5
0.12
3
CBMV_ROBtwo views11.77
17
2.89
18
6.92
7
5.20
2
2.88
14
14.12
9
4.79
8
26.69
11
29.84
19
26.17
33
31.92
23
23.61
15
20.74
39
20.20
25
10.18
13
1.78
54
1.80
66
2.17
59
1.26
53
1.61
35
0.55
18
iResNetv2_ROBtwo views12.00
18
2.91
19
13.46
27
5.82
6
3.73
21
13.15
6
6.45
19
34.15
39
36.02
35
25.25
30
33.81
32
29.79
30
14.24
18
13.61
5
6.10
4
0.21
13
0.02
5
0.12
11
0.00
1
0.82
21
0.25
9
NLCA_NET_v2_RVCtwo views12.05
19
3.43
27
16.25
43
7.62
24
5.58
38
14.57
13
6.00
17
32.98
36
27.41
13
21.69
17
31.40
21
25.21
22
13.47
12
14.96
10
15.60
24
0.84
33
0.33
32
0.35
21
0.35
35
1.29
31
1.78
40
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-ROBtwo views12.13
20
3.51
28
15.82
39
7.52
23
5.50
37
14.30
11
6.60
20
32.79
34
27.59
14
21.97
19
32.14
25
24.76
21
14.02
15
14.82
8
16.06
25
0.85
35
0.35
34
0.34
20
0.35
35
1.30
33
1.94
42
SGM-Foresttwo views12.92
21
1.87
6
6.61
3
5.68
5
2.05
4
23.19
38
11.30
35
34.24
40
30.77
22
26.54
34
31.99
24
29.98
31
15.74
22
21.17
29
13.36
21
1.05
43
0.33
32
0.84
38
0.01
10
0.71
17
0.97
24
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
TDLMtwo views12.98
22
3.78
32
9.77
19
10.75
44
4.94
32
16.10
19
14.08
42
35.05
45
31.87
25
22.36
20
26.51
8
26.09
25
14.97
20
24.73
38
13.11
20
1.05
43
0.05
10
1.21
50
0.27
30
1.83
38
1.00
26
AANet_RVCtwo views13.16
23
5.34
43
10.83
22
8.20
29
4.44
27
14.13
10
9.46
27
28.78
17
37.67
41
23.44
25
37.47
41
23.48
13
15.83
24
22.39
30
16.63
28
1.67
52
0.86
56
0.24
17
0.02
11
0.63
16
1.60
37
CVANet_RVCtwo views13.24
24
3.39
26
8.43
14
8.42
30
5.04
34
18.52
23
11.72
37
32.03
30
33.85
29
23.17
24
32.77
26
29.78
29
16.51
25
23.20
31
11.28
19
0.94
38
0.04
8
1.19
49
0.47
39
3.13
51
1.01
27
DLCB_ROBtwo views13.35
25
3.00
22
9.12
16
9.43
35
5.68
39
21.80
34
10.12
29
29.19
18
29.92
20
27.71
37
31.45
22
32.37
35
19.50
34
19.02
21
16.73
29
0.22
14
0.04
8
0.38
23
0.06
19
0.49
10
0.85
21
StereoDRNet-Refinedtwo views13.36
26
2.80
17
10.82
21
7.94
28
3.10
16
18.95
24
4.45
6
28.47
16
29.65
18
29.11
41
41.47
55
24.44
19
17.64
30
23.34
32
21.66
34
0.16
10
0.07
17
0.53
31
0.32
33
0.77
20
1.64
38
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
CBMVpermissivetwo views13.69
27
3.63
29
8.02
12
5.93
7
2.69
11
22.57
35
12.44
39
29.81
20
31.57
24
31.20
50
33.79
31
31.04
33
17.41
29
25.28
40
14.11
22
0.71
30
0.60
49
0.60
32
0.11
22
1.11
27
1.13
29
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
pmcnntwo views14.59
28
3.68
31
19.78
53
6.83
19
4.51
28
21.27
32
14.67
45
27.37
12
30.54
21
27.09
36
40.19
50
38.71
53
18.53
33
17.70
19
19.27
30
0.84
33
0.09
22
0.00
1
0.00
1
0.48
7
0.31
13
NVstereo2Dtwo views15.15
29
2.97
21
16.28
44
9.23
33
7.56
53
30.03
54
9.53
28
42.80
72
42.72
51
15.11
2
27.03
9
23.16
11
16.93
27
24.28
35
15.45
23
3.07
67
0.43
41
1.44
51
0.61
43
6.74
72
7.60
77
StereoDRNettwo views15.30
30
4.47
37
14.35
33
10.71
43
8.69
58
25.02
44
11.03
34
39.66
62
35.62
33
29.17
42
28.81
13
34.05
42
17.80
31
18.80
20
23.77
42
0.41
24
0.30
31
0.15
13
0.14
23
1.64
36
1.42
33
DRN-Testtwo views15.88
31
3.83
34
14.70
34
9.96
39
8.02
56
28.93
53
14.65
44
42.34
68
38.25
44
28.18
40
34.31
35
29.22
28
18.35
32
20.01
24
22.68
38
0.37
20
0.41
37
0.46
28
0.28
31
1.28
30
1.34
32
PA-Nettwo views16.29
32
6.88
54
27.34
61
9.88
37
11.12
66
21.77
33
28.15
78
32.06
31
39.49
45
19.99
12
27.91
11
23.50
14
20.59
38
20.25
26
24.04
45
0.24
16
2.26
74
0.22
15
4.59
77
1.13
28
4.48
68
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
DISCOtwo views16.36
33
1.88
7
16.15
41
7.78
27
4.19
24
26.92
47
10.21
31
30.31
22
44.26
57
21.35
16
33.91
33
35.80
47
22.40
41
35.65
68
33.87
67
0.18
11
0.08
19
0.04
9
0.04
15
1.71
37
0.54
17
NaN_ROBtwo views16.51
34
6.51
49
16.22
42
10.53
42
4.64
30
31.76
60
17.78
55
37.00
53
43.37
54
29.68
46
31.29
19
32.67
39
20.32
37
28.00
46
16.06
25
0.38
21
0.41
37
0.32
19
0.41
37
0.87
22
1.94
42
DANettwo views16.85
35
8.00
57
26.04
60
14.56
62
7.25
51
20.59
28
5.37
10
26.22
10
26.02
12
32.17
56
36.00
36
37.20
49
29.77
60
30.75
49
24.62
49
1.14
47
1.34
60
2.21
60
0.51
41
3.18
52
4.07
64
PSMNet_ROBtwo views16.90
36
5.45
44
14.16
31
13.77
57
7.25
51
27.65
49
32.74
82
43.21
74
37.88
42
24.00
26
32.97
28
34.55
44
16.69
26
17.41
17
23.96
44
0.38
21
0.22
26
0.93
40
2.05
63
1.83
38
0.96
23
GANettwo views16.92
37
4.32
36
12.46
24
10.84
45
4.24
25
23.04
37
15.36
48
39.91
63
34.47
31
32.42
57
45.10
67
40.04
59
27.54
55
23.61
34
20.20
31
1.02
42
0.08
19
0.67
35
0.22
26
1.87
40
0.97
24
NCCL2two views17.23
38
6.37
47
14.30
32
23.45
82
7.18
50
24.16
41
17.05
53
34.39
42
25.88
11
31.55
51
38.01
43
42.09
64
26.81
52
20.94
28
22.69
39
0.45
26
0.23
27
2.63
64
1.80
59
2.02
42
2.57
49
ADCReftwo views17.30
39
6.73
52
42.01
75
10.11
40
8.78
59
26.61
46
10.45
32
31.29
25
32.75
28
31.19
49
41.62
57
18.47
7
19.78
36
24.29
36
34.06
69
0.76
32
0.42
40
2.16
58
1.79
58
1.23
29
1.56
36
XPNet_ROBtwo views17.33
40
4.84
41
15.44
37
11.15
48
6.64
45
20.88
29
16.10
50
36.26
48
39.54
46
31.80
54
41.32
54
41.30
62
23.91
42
24.67
37
27.40
56
1.07
46
0.74
53
0.79
37
0.25
28
1.10
26
1.31
31
RPtwo views17.58
41
4.92
42
15.27
36
15.44
64
11.95
71
21.05
30
11.75
38
27.64
13
45.09
60
22.99
23
42.05
60
39.39
56
27.74
56
25.54
41
22.33
37
5.45
76
0.59
48
4.33
74
1.62
57
3.53
56
2.86
52
ETE_ROBtwo views17.67
42
9.48
62
17.66
47
13.60
55
4.57
29
23.03
36
21.09
63
32.49
33
35.24
32
30.49
47
38.06
44
46.18
74
24.49
44
26.40
43
24.21
46
0.46
27
0.24
28
1.14
46
0.84
49
1.52
34
2.13
46
PWCDC_ROBbinarytwo views17.80
43
9.62
63
18.38
49
17.95
70
7.74
54
24.61
42
5.56
14
34.41
43
52.79
74
29.66
45
50.20
75
20.63
9
19.69
35
29.85
48
20.87
32
6.18
77
0.26
30
0.10
10
0.05
16
5.36
69
2.00
44
MDST_ROBtwo views17.92
44
1.60
4
13.68
28
13.88
58
6.44
43
43.05
79
14.86
46
42.74
71
41.66
50
43.25
75
42.85
62
28.43
26
17.28
28
29.35
47
16.06
25
0.88
36
0.11
24
0.63
34
0.47
39
0.50
11
0.64
19
ADCP+two views18.16
45
4.61
39
32.94
68
9.31
34
10.23
63
28.87
52
11.69
36
33.43
37
36.49
36
29.28
44
40.49
51
24.18
17
24.61
46
33.89
60
35.33
70
0.25
17
0.40
36
2.00
55
1.17
51
2.16
44
1.87
41
PWC_ROBbinarytwo views18.37
46
10.55
67
25.19
59
10.21
41
6.29
42
23.75
39
4.79
8
35.93
47
48.75
64
32.56
58
44.46
66
34.68
45
24.78
47
30.99
50
25.72
53
1.38
50
0.08
19
1.57
52
0.26
29
2.18
45
3.17
55
Anonymous Stereotwo views18.60
47
11.75
72
49.81
82
14.43
60
12.02
72
14.33
12
23.23
71
32.16
32
43.08
53
24.32
28
34.16
34
24.24
18
14.12
16
31.67
53
30.84
63
0.89
37
0.91
57
1.74
53
1.92
62
3.23
53
3.23
56
stereogantwo views18.62
48
3.07
23
16.31
45
13.04
54
9.99
62
35.74
70
9.09
26
38.17
60
45.27
61
24.39
29
41.24
53
39.93
58
25.46
49
31.29
52
24.33
48
1.80
55
0.91
57
2.71
66
0.75
48
5.15
68
3.70
58
RYNettwo views18.92
49
4.69
40
16.88
46
10.99
46
15.84
79
46.72
81
16.57
52
37.84
57
48.03
62
26.92
35
26.48
6
37.84
52
24.92
48
19.80
23
31.04
64
0.25
17
0.25
29
0.62
33
0.03
14
6.34
71
6.39
74
LALA_ROBtwo views19.19
50
7.15
56
15.93
40
12.96
53
5.30
36
28.30
51
23.21
70
42.51
69
36.69
37
33.45
61
40.81
52
50.96
80
24.53
45
27.92
45
25.66
52
0.60
28
0.44
42
2.04
56
1.22
52
2.54
47
1.47
34
SGM_RVCbinarytwo views19.52
51
2.12
9
6.79
5
6.59
16
1.33
2
38.20
76
15.74
49
38.13
59
34.07
30
45.71
79
41.91
59
51.41
81
38.10
74
38.77
75
27.80
59
0.61
29
0.36
35
0.51
30
0.33
34
1.03
24
0.90
22
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
RTSCtwo views19.64
52
9.36
59
31.06
66
10.99
46
6.24
41
28.10
50
10.14
30
37.50
56
58.34
85
31.74
53
36.70
39
32.50
36
21.03
40
36.00
69
37.24
75
1.23
48
0.44
42
0.12
11
0.31
32
1.98
41
1.69
39
NCC-stereotwo views19.75
53
6.08
45
23.96
57
15.05
63
16.17
80
18.42
22
21.45
64
43.31
75
41.00
49
24.30
27
38.12
45
34.98
46
26.92
53
35.46
66
25.65
51
4.29
72
2.07
70
3.24
70
8.90
84
2.86
49
2.86
52
RGCtwo views19.90
54
13.26
73
20.24
54
18.19
71
14.06
77
21.21
31
14.33
43
34.97
44
43.60
56
27.73
38
41.49
56
39.76
57
27.98
57
34.52
63
21.37
33
3.66
69
0.54
47
8.67
81
4.16
75
4.19
61
4.02
62
DeepPrunerFtwo views19.92
55
11.02
71
44.29
76
20.74
75
17.75
82
19.19
25
22.57
65
36.86
52
49.93
67
25.48
31
36.62
38
24.69
20
23.97
43
23.44
33
21.79
35
2.68
66
1.63
65
5.70
76
3.95
74
3.38
54
2.67
51
WCMA_ROBtwo views20.02
56
4.30
35
19.72
52
9.47
36
7.00
49
32.71
65
13.99
41
31.97
29
32.48
27
41.51
73
52.00
77
44.09
67
36.14
72
32.43
55
24.29
47
6.19
78
2.79
78
1.09
45
1.10
50
3.95
57
3.13
54
G-Nettwo views20.18
57
6.57
50
18.57
50
15.59
65
17.87
83
24.95
43
23.05
67
44.94
78
40.82
48
22.84
21
41.69
58
37.44
50
25.78
50
33.88
58
26.23
54
5.26
75
1.82
67
3.47
72
6.14
80
3.48
55
3.26
57
SANettwo views20.96
58
6.60
51
29.81
64
8.83
31
3.19
17
31.27
59
20.56
62
41.86
67
56.09
82
39.30
69
43.62
63
44.95
70
31.93
66
27.83
44
23.67
41
0.94
38
0.52
46
0.99
42
0.42
38
4.72
63
2.03
45
FBW_ROBtwo views21.00
59
10.23
66
22.72
55
13.71
56
6.77
46
30.49
58
16.46
51
44.58
76
43.57
55
44.25
78
39.59
48
43.27
66
26.12
51
33.96
61
21.80
36
2.58
65
1.84
68
7.14
80
2.78
67
4.00
59
4.06
63
SHDtwo views21.32
60
10.03
65
30.16
65
14.30
59
8.68
57
24.15
40
8.93
24
40.82
65
61.17
87
35.79
64
44.15
65
38.73
54
30.20
61
33.88
58
31.09
65
2.00
56
0.81
54
1.17
47
1.55
55
4.08
60
4.78
69
CSANtwo views21.34
61
9.45
61
23.34
56
20.99
77
4.95
33
32.57
63
34.26
83
38.83
61
49.95
69
36.97
66
39.72
49
44.97
71
31.73
65
26.05
42
23.94
43
1.52
51
0.47
44
0.85
39
1.43
54
2.73
48
2.14
48
ADCLtwo views21.64
62
6.20
46
47.32
80
9.93
38
6.91
48
38.69
77
19.97
61
31.26
24
54.04
76
27.89
39
47.97
72
31.35
34
30.66
63
33.18
57
36.62
72
0.99
40
0.92
59
2.63
64
1.82
60
2.45
46
2.13
46
ADCPNettwo views21.93
63
9.38
60
57.92
85
11.76
51
6.88
47
36.03
71
18.44
56
32.80
35
35.93
34
32.91
60
43.80
64
39.21
55
26.99
54
31.14
51
36.94
74
2.05
57
2.67
77
2.30
62
3.45
70
4.22
62
3.82
59
MeshStereopermissivetwo views22.27
64
6.87
53
11.15
23
7.69
25
4.87
31
37.70
75
13.06
40
41.64
66
36.95
38
50.92
83
53.41
78
58.12
84
41.93
80
37.73
72
27.62
58
2.52
64
2.37
75
2.99
68
2.07
64
3.09
50
2.60
50
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
ADCMidtwo views22.76
65
10.75
68
41.73
74
11.23
50
7.97
55
27.26
48
19.18
60
35.44
46
38.02
43
40.23
70
46.38
70
36.56
48
42.70
81
38.20
74
40.75
80
1.77
53
1.43
62
3.05
69
3.81
73
4.96
65
3.82
59
MSMD_ROBtwo views22.90
66
9.64
64
14.76
35
17.46
69
9.88
61
34.94
68
18.83
57
33.67
38
37.47
39
40.97
72
59.80
83
45.93
72
38.34
75
31.74
54
23.24
40
6.22
79
3.58
82
8.78
82
11.03
85
6.76
73
5.06
70
LE_ROBtwo views22.93
67
3.07
23
14.02
29
7.70
26
2.86
13
31.99
62
17.35
54
28.10
15
67.19
93
70.51
94
55.61
82
49.07
78
47.62
85
24.76
39
36.49
71
0.35
19
0.20
25
0.27
18
0.55
42
0.48
7
0.44
16
AnyNet_C32two views23.37
68
13.28
74
40.29
73
12.36
52
11.44
68
30.39
55
29.15
79
31.36
26
44.90
59
35.36
62
48.86
73
33.61
40
34.20
69
43.15
79
41.62
83
1.27
49
1.35
61
1.18
48
2.58
66
4.77
64
6.27
73
SGM-ForestMtwo views23.72
69
2.44
15
9.13
17
7.43
21
2.25
6
44.80
80
19.11
59
44.90
77
49.97
70
50.74
82
51.18
76
62.06
90
45.83
84
44.96
80
33.88
68
1.00
41
0.84
55
0.78
36
0.62
44
1.29
31
1.11
28
XQCtwo views24.10
70
16.76
77
50.68
83
21.37
79
11.01
65
35.24
69
18.84
58
37.28
55
55.11
79
31.64
52
30.06
17
37.71
51
30.31
62
37.17
70
39.66
78
4.20
71
0.41
37
2.76
67
1.89
61
11.46
83
8.42
81
FC-DCNNcopylefttwo views24.37
71
10.99
70
19.02
51
18.44
72
9.16
60
36.98
73
23.07
68
40.44
64
43.05
52
46.25
81
53.58
79
50.44
79
37.86
73
35.45
65
27.57
57
6.87
81
3.37
79
6.01
77
4.88
78
7.86
75
6.05
71
DPSNettwo views24.40
72
6.97
55
33.14
69
11.16
49
6.54
44
53.33
85
43.32
89
51.28
86
59.37
86
30.89
48
39.36
46
40.35
60
34.30
70
32.57
56
25.09
50
4.65
74
1.62
64
0.37
22
0.66
45
8.51
77
4.44
67
PDISCO_ROBtwo views24.45
73
9.24
58
29.28
63
28.68
85
19.96
84
37.14
74
15.33
47
45.04
79
54.69
78
29.24
43
42.67
61
46.07
73
28.17
58
35.35
64
30.30
62
9.92
84
2.13
71
6.90
78
3.20
68
8.74
78
6.87
76
Nwc_Nettwo views25.10
74
6.41
48
28.02
62
21.66
80
13.13
73
41.14
78
24.19
72
52.57
88
52.84
75
32.10
55
47.48
71
44.22
68
32.86
67
48.01
83
43.12
85
4.50
73
0.73
52
0.96
41
0.15
24
3.98
58
4.00
61
GANetREF_RVCpermissivetwo views25.41
75
29.28
87
24.63
58
25.18
83
6.04
40
30.44
56
35.39
85
47.84
80
50.60
72
36.88
65
36.04
37
34.11
43
31.23
64
37.46
71
27.96
61
8.58
82
2.66
76
16.29
86
4.95
79
14.37
87
8.32
80
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
ADCStwo views26.20
76
13.71
75
46.90
79
14.51
61
10.34
64
30.46
57
22.72
66
42.81
73
57.61
84
42.58
74
49.37
74
41.91
63
40.43
79
41.90
78
44.58
87
2.37
61
2.19
73
2.37
63
3.29
69
7.32
74
6.52
75
Abc-Nettwo views27.90
77
15.61
76
44.53
77
32.29
87
15.22
78
36.24
72
29.76
80
50.37
85
50.13
71
39.09
68
60.11
84
46.20
75
39.95
78
37.96
73
32.83
66
6.22
79
1.43
62
4.06
73
3.61
71
6.26
70
6.07
72
PASMtwo views28.19
78
17.35
78
48.18
81
22.49
81
22.20
85
25.92
45
25.84
75
34.35
41
49.45
65
35.37
63
39.48
47
46.72
76
33.36
68
34.28
62
36.77
73
11.73
87
14.94
88
17.43
88
22.18
89
13.45
84
12.32
85
AnyNet_C01two views29.76
79
25.78
86
75.37
91
17.03
68
11.34
67
49.00
83
27.12
77
37.99
58
40.05
47
40.73
71
65.13
91
46.98
77
38.86
77
46.63
81
48.19
88
2.14
58
2.13
71
2.13
57
2.56
65
8.14
76
7.98
79
LSMtwo views29.98
80
19.52
83
55.11
84
19.06
73
53.75
92
32.57
63
23.17
69
42.61
70
48.57
63
45.97
80
55.04
81
42.93
65
36.02
71
35.63
67
27.03
55
2.33
59
9.19
87
4.87
75
7.71
83
10.65
82
27.91
90
RTStwo views30.15
81
23.12
84
96.33
95
16.71
66
11.63
69
58.66
87
24.28
73
36.80
50
62.63
89
43.79
76
46.10
68
32.56
37
43.90
82
51.29
85
40.92
81
2.38
62
0.66
50
1.06
43
0.74
46
5.11
66
4.34
65
RTSAtwo views30.15
81
23.12
84
96.33
95
16.71
66
11.63
69
58.66
87
24.28
73
36.80
50
62.63
89
43.79
76
46.10
68
32.56
37
43.90
82
51.29
85
40.92
81
2.38
62
0.66
50
1.06
43
0.74
46
5.11
66
4.34
65
DispFullNettwo views30.27
83
31.74
88
58.63
86
30.13
86
28.18
87
31.79
61
8.94
25
37.16
54
54.68
77
32.60
59
37.99
42
44.62
69
28.80
59
41.55
77
27.93
60
11.94
88
3.46
80
35.53
90
12.34
86
30.08
89
17.26
86
MANEtwo views31.97
84
10.75
68
17.71
48
20.94
76
13.77
75
47.80
82
26.85
76
53.00
89
55.20
80
59.29
86
64.30
89
64.93
92
56.31
90
51.83
88
38.69
77
9.61
83
4.72
83
6.91
79
19.93
88
9.20
79
7.64
78
PVDtwo views33.19
85
18.59
80
37.64
72
26.09
84
16.99
81
33.25
66
42.25
88
51.42
87
66.03
91
52.65
85
64.31
90
51.93
82
55.13
88
51.61
87
40.40
79
4.00
70
3.51
81
9.53
83
4.26
76
10.62
81
23.56
87
ELAScopylefttwo views33.68
86
18.95
81
36.71
71
20.64
74
13.61
74
56.43
86
35.14
84
50.03
84
49.90
66
60.58
87
63.83
88
58.64
86
53.36
86
50.73
84
44.44
86
11.03
85
6.58
85
10.02
84
7.43
82
13.58
86
11.94
84
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ELAS_RVCcopylefttwo views33.79
87
19.15
82
35.97
70
21.28
78
13.79
76
52.58
84
36.11
86
48.96
83
55.66
81
61.47
90
62.91
86
58.35
85
53.55
87
53.37
89
42.23
84
11.25
86
6.64
86
10.32
85
7.05
81
13.56
85
11.58
83
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SAMSARAtwo views34.43
88
18.53
79
31.91
67
55.34
92
35.34
89
75.99
93
94.71
97
47.97
81
50.84
73
37.47
67
36.99
40
53.12
83
38.55
76
39.56
76
38.65
76
3.59
68
5.40
84
2.22
61
3.72
72
9.43
80
9.29
82
NVStereoNet_ROBtwo views41.93
89
32.40
89
44.55
78
36.60
88
27.78
86
34.52
67
32.38
81
48.84
82
57.18
83
69.34
93
75.62
95
66.07
93
64.86
94
47.34
82
60.78
91
15.72
89
23.99
89
23.40
89
32.06
90
21.34
88
23.86
88
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+two views53.31
90
68.61
93
98.75
97
67.03
94
37.58
90
79.83
95
81.74
95
60.38
92
61.61
88
51.43
84
54.92
80
61.11
89
59.29
92
69.66
93
76.71
95
23.88
90
24.03
90
17.08
87
12.35
87
32.47
90
27.69
89
SGM+DAISYtwo views54.67
91
56.61
90
62.73
88
40.21
89
53.22
91
61.95
90
48.59
91
55.05
90
44.62
58
61.03
89
63.68
87
60.46
88
55.62
89
60.68
91
55.34
89
56.34
92
48.56
94
50.16
91
49.63
93
52.25
91
56.76
93
SPS-STEREOcopylefttwo views55.62
92
59.14
91
64.16
89
45.05
90
53.84
93
59.88
89
44.00
90
59.53
91
49.94
68
62.33
91
61.09
85
59.80
87
56.82
91
60.06
90
57.85
90
58.41
93
48.34
93
51.07
92
49.21
92
55.41
92
56.48
92
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
PWCKtwo views65.90
93
88.63
96
79.40
93
80.32
96
29.96
88
63.85
91
39.45
87
72.07
94
71.26
94
78.74
95
68.99
92
77.82
95
61.77
93
82.81
96
62.64
93
80.17
94
39.88
91
81.90
96
47.37
91
73.54
95
37.51
91
edge stereotwo views68.62
94
66.23
92
75.79
92
66.56
93
62.76
94
80.29
96
72.62
93
72.56
95
75.42
95
68.91
92
80.34
96
71.80
94
67.18
95
70.37
94
75.18
94
53.98
91
47.53
92
59.61
93
61.48
94
69.05
93
74.78
94
MFMNet_retwo views72.05
95
79.64
94
64.23
90
52.73
91
84.71
96
77.93
94
68.35
92
63.96
93
66.93
92
61.00
88
71.29
93
63.84
91
68.91
96
72.42
95
61.11
92
82.03
96
74.22
96
80.34
95
74.44
95
83.62
96
89.34
96
DPSimNet_ROBtwo views74.29
96
83.27
95
59.30
87
77.19
95
65.13
95
75.14
92
74.95
94
73.99
96
76.94
96
81.99
96
75.48
94
80.18
96
71.76
97
67.39
92
84.00
96
81.87
95
51.14
95
79.99
94
76.70
96
70.17
94
79.21
95
MADNet++two views94.16
97
94.02
97
90.25
94
96.28
97
96.84
97
97.17
97
90.16
96
94.71
97
91.69
97
97.07
97
93.71
97
94.62
97
92.63
98
96.15
97
95.18
97
95.42
97
95.89
97
92.80
97
92.23
97
92.10
97
94.30
97
MEDIAN_ROBtwo views99.19
98
99.84
98
99.62
99
98.49
98
98.51
98
98.58
98
97.81
99
98.80
98
98.56
98
99.36
100
99.49
98
99.56
98
99.06
99
98.35
98
98.31
98
99.99
98
99.63
98
100.00
98
100.00
98
99.81
98
99.95
98
AVERAGE_ROBtwo views99.80
99
99.99
99
99.40
98
100.00
99
100.00
99
98.99
99
97.72
98
100.00
99
100.00
99
100.00
101
100.00
99
100.00
101
100.00
100
100.00
101
100.00
101
100.00
99
100.00
99
100.00
98
100.00
98
100.00
99
100.00
99
DGTPSM_ROBtwo views99.95
100
100.00
100
100.00
100
100.00
99
100.00
99
100.00
100
100.00
100
100.00
99
100.00
99
99.14
98
100.00
99
99.96
99
100.00
100
99.99
99
99.99
99
100.00
99
100.00
99
100.00
98
100.00
98
100.00
99
100.00
99
DPSMNet_ROBtwo views99.95
100
100.00
100
100.00
100
100.00
99
100.00
99
100.00
100
100.00
100
100.00
99
100.00
99
99.15
99
100.00
99
99.96
99
100.00
100
99.99
99
99.99
99
100.00
99
100.00
99
100.00
98
100.00
98
100.00
99
100.00
99
DPSMtwo views100.00
102
100.00
100
100.00
100
100.00
99
100.00
99
100.00
100
100.00
100
100.00
99
100.00
99
100.00
101
100.00
99
100.00
101
100.00
100
100.00
101
100.00
101
100.00
99
100.00
99
100.00
98
100.00
98
100.00
99
100.00
99
LSM0two views100.00
102
100.00
100
100.00
100
100.00
99
100.00
99
100.00
100
100.00
100
100.00
99
100.00
99
100.00
101
100.00
99
100.00
101
100.00
100
100.00
101
100.00
101
100.00
99
100.00
99
100.00
98
100.00
98
100.00
99
100.00
99
DPSM_ROBtwo views100.00
102
100.00
100
100.00
100
100.00
99
100.00
99
100.00
100
100.00
100
100.00
99
100.00
99
100.00
101
100.00
99
100.00
101
100.00
100
100.00
101
100.00
101
100.00
99
100.00
99
100.00
98
100.00
98
100.00
99
100.00
99
MSMDNettwo views7.19
3