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 views1.20
1
0.58
2
1.81
1
0.99
2
0.67
4
3.78
37
0.90
13
4.09
49
1.51
1
2.16
17
1.71
8
1.13
1
0.94
1
1.40
1
0.91
1
0.30
2
0.24
10
0.26
2
0.21
3
0.26
1
0.25
1
R-Stereotwo views1.20
1
0.58
2
1.81
1
0.99
2
0.67
4
3.78
37
0.90
13
4.09
49
1.51
1
2.16
17
1.71
8
1.13
1
0.94
1
1.40
1
0.91
1
0.30
2
0.24
10
0.26
2
0.21
3
0.26
1
0.25
1
AdaStereotwo views1.45
3
0.83
14
2.34
5
1.28
27
0.81
10
3.21
5
1.14
24
4.20
58
2.46
33
2.05
13
1.77
13
2.48
18
1.18
5
1.71
3
1.44
3
0.43
27
0.25
13
0.43
27
0.28
12
0.44
15
0.34
10
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
StereoDRNet-Refinedtwo views1.77
4
0.81
11
2.22
4
1.21
19
0.76
7
3.26
9
0.71
4
3.82
18
2.05
6
2.29
21
2.05
30
2.98
37
1.44
14
7.15
32
2.27
30
0.30
2
0.23
8
0.42
26
0.33
15
0.45
19
0.64
45
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
MLCVtwo views1.87
5
0.91
17
8.59
16
1.10
10
0.57
2
3.59
26
0.53
1
3.70
13
2.20
10
2.80
42
2.25
44
2.00
7
1.42
13
4.13
14
1.86
22
0.30
2
0.20
3
0.32
9
0.21
3
0.33
6
0.36
13
iResNet_ROBtwo views1.89
6
1.02
25
8.60
17
1.39
45
0.98
25
3.72
35
0.66
3
3.84
20
2.65
39
1.88
7
2.05
30
1.98
6
1.27
8
4.52
16
1.67
13
0.30
2
0.20
3
0.28
5
0.18
1
0.31
4
0.40
15
CFNet_RVCtwo views1.90
7
0.95
21
3.56
11
1.07
6
1.54
57
4.09
60
0.90
13
2.85
1
1.82
5
1.78
3
1.70
5
2.75
26
1.48
18
9.33
44
1.72
16
0.33
11
0.30
19
0.61
62
0.36
23
0.52
26
0.39
14
DN-CSS_ROBtwo views1.91
8
1.55
61
11.13
27
1.82
73
0.88
15
3.99
56
0.61
2
4.01
39
1.67
3
1.64
2
1.89
20
1.39
4
1.06
3
3.23
6
1.50
4
0.31
7
0.19
2
0.44
29
0.29
13
0.37
8
0.31
7
LALA_ROBtwo views1.95
9
1.48
58
3.11
9
1.05
5
0.98
25
3.99
56
1.82
51
4.44
74
2.28
18
3.31
56
1.85
18
2.76
29
1.67
27
4.04
13
3.23
44
0.46
32
0.41
37
0.55
53
0.52
52
0.60
41
0.57
34
DeepPruner_ROBtwo views2.05
10
1.09
33
10.50
24
1.01
4
1.20
37
4.39
70
0.91
17
3.14
3
1.72
4
2.60
30
1.54
1
2.78
30
1.27
8
4.52
16
1.72
16
0.51
42
0.40
35
0.37
13
0.34
16
0.48
21
0.52
28
DLCB_ROBtwo views2.05
10
0.95
21
2.37
6
1.24
22
1.02
29
2.92
3
1.30
29
3.68
11
2.30
20
2.91
45
2.17
38
3.32
42
1.52
20
10.37
53
2.50
32
0.40
23
0.33
22
0.40
22
0.36
23
0.41
12
0.47
21
ETE_ROBtwo views2.07
12
1.40
54
3.01
8
1.09
7
0.92
20
3.94
49
1.53
40
4.25
61
2.23
12
3.31
56
2.06
32
2.35
14
1.52
20
8.74
40
2.07
25
0.44
29
0.36
30
0.51
43
0.48
49
0.54
30
0.62
41
HITNettwo views2.07
12
0.82
13
14.77
42
1.34
36
0.56
1
3.47
19
0.80
7
4.07
47
2.35
24
2.16
17
1.61
2
1.64
5
1.44
14
2.09
4
2.56
33
0.26
1
0.18
1
0.41
25
0.26
11
0.45
19
0.25
1
DISCOtwo views2.25
14
0.73
8
10.06
22
1.76
71
1.12
32
3.47
19
1.25
28
3.24
4
2.26
16
2.65
32
1.96
22
4.53
69
1.92
37
4.17
15
3.61
47
0.32
9
0.25
13
0.31
8
0.29
13
0.66
51
0.41
17
NOSS_ROBtwo views2.27
15
0.73
8
3.00
7
1.71
68
1.00
27
3.90
46
0.97
18
4.29
64
2.81
49
2.94
46
2.19
40
3.88
53
1.19
6
11.77
73
1.62
8
0.59
53
0.54
60
0.56
54
0.52
52
0.56
36
0.54
31
XPNet_ROBtwo views2.27
15
1.11
36
4.10
12
1.36
41
0.97
23
3.94
49
1.63
44
4.07
47
2.14
8
2.82
43
2.00
25
2.32
13
1.73
29
10.81
58
3.39
46
0.51
42
0.47
52
0.48
35
0.40
38
0.51
24
0.56
33
TDLMtwo views2.29
17
1.08
31
4.61
13
1.29
29
1.01
28
4.26
68
2.17
60
3.75
16
2.79
45
2.31
23
1.98
24
2.06
10
1.59
22
12.45
80
1.72
16
0.51
42
0.28
18
0.52
49
0.35
19
0.56
36
0.50
25
CFNettwo views2.34
18
1.10
34
8.96
18
1.34
36
1.24
39
4.09
60
0.73
6
4.04
43
2.77
43
1.62
1
1.87
19
2.27
12
1.45
17
11.41
63
1.82
21
0.31
7
0.24
10
0.43
27
0.36
23
0.43
14
0.33
8
HSMtwo views2.35
19
0.81
11
2.21
3
1.14
13
1.55
59
3.44
15
1.03
19
3.99
38
2.25
14
2.29
21
2.11
35
8.04
83
3.05
58
11.60
67
1.71
14
0.33
11
0.22
7
0.28
5
0.23
8
0.32
5
0.35
12
iResNettwo views2.52
20
0.94
20
21.17
87
1.84
75
0.72
6
3.75
36
0.81
9
4.14
54
2.55
38
2.33
24
2.03
26
1.35
3
1.44
14
3.27
7
2.16
28
0.35
18
0.23
8
0.28
5
0.22
7
0.44
15
0.45
19
NCCL2two views2.53
21
1.35
48
10.90
25
1.15
14
0.97
23
3.53
24
2.56
65
3.65
10
2.23
12
2.76
38
1.80
15
2.41
16
1.66
26
10.35
52
2.07
25
0.44
29
0.40
35
0.57
55
0.53
55
0.59
38
0.65
47
NVstereo2Dtwo views2.54
22
0.90
15
11.10
26
1.32
31
1.47
54
3.58
25
1.38
32
4.13
52
2.45
31
1.90
9
1.71
8
2.86
34
1.76
30
10.86
59
1.66
12
0.60
54
0.43
43
0.53
50
0.41
39
0.78
61
0.94
72
ccs_robtwo views2.54
22
1.14
38
14.21
38
1.35
39
0.96
21
3.98
55
0.71
4
3.86
23
2.88
50
2.04
12
2.07
33
2.05
9
1.29
11
10.60
55
1.71
14
0.32
9
0.25
13
0.38
15
0.35
19
0.39
9
0.33
8
NLCA_NET_v2_RVCtwo views2.54
22
1.04
27
16.73
58
1.53
56
2.00
83
3.96
52
0.87
12
3.87
25
2.31
21
2.54
28
1.67
4
2.85
33
1.63
24
5.55
23
1.62
8
0.46
32
0.33
22
0.38
15
0.38
30
0.53
29
0.62
41
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
CVANet_RVCtwo views2.55
25
1.10
34
9.13
19
1.46
47
1.15
34
3.96
52
1.53
40
3.91
33
2.50
35
3.04
48
2.04
28
2.60
23
1.72
28
12.46
81
1.50
4
0.49
37
0.30
19
0.51
43
0.41
39
0.63
43
0.50
25
CC-Net-ROBtwo views2.55
25
1.05
28
16.77
59
1.50
52
1.94
82
3.97
54
0.90
13
3.85
21
2.25
14
2.55
29
1.70
5
2.82
32
1.65
25
5.67
26
1.65
10
0.47
35
0.33
22
0.38
15
0.39
33
0.54
30
0.64
45
DANettwo views2.63
27
1.07
30
8.10
14
1.83
74
1.70
71
3.25
8
1.16
25
4.31
69
2.38
28
3.40
59
2.18
39
4.08
59
1.80
32
9.77
46
3.89
50
0.52
47
0.55
63
0.59
58
0.44
43
0.63
43
0.86
68
iResNetv2_ROBtwo views2.65
28
5.00
87
19.28
76
2.06
85
1.09
31
3.83
43
1.13
23
4.14
54
2.09
7
1.85
5
2.03
26
2.01
8
1.51
19
3.58
9
1.58
6
0.34
14
0.21
5
0.32
9
0.21
3
0.44
15
0.29
5
PASMtwo views2.67
29
2.05
76
9.71
21
1.73
70
1.37
49
2.53
1
1.52
39
3.10
2
2.17
9
2.79
41
1.84
16
2.90
35
1.85
35
11.89
76
2.19
29
0.81
74
0.85
83
0.87
83
1.01
86
1.14
80
1.05
75
PWC_ROBbinarytwo views2.69
30
2.01
74
8.36
15
1.33
33
1.90
80
3.23
6
1.18
26
4.33
71
3.25
65
3.11
51
3.41
68
2.71
24
2.18
41
9.15
42
4.48
52
0.56
51
0.31
21
0.61
62
0.36
23
0.69
53
0.74
56
RPtwo views2.79
31
1.20
41
10.36
23
1.54
58
1.93
81
3.44
15
1.97
54
4.05
45
2.34
23
2.66
33
2.79
54
3.98
56
2.58
49
10.21
51
2.61
34
0.95
80
0.47
52
0.71
73
0.57
61
0.71
57
0.68
51
SHDtwo views2.82
32
2.04
75
13.80
34
1.69
67
1.79
77
2.73
2
1.07
21
3.85
21
3.18
60
4.63
74
2.74
53
3.38
43
2.77
53
5.48
22
3.66
48
0.62
57
0.47
52
0.51
43
0.55
57
0.65
48
0.74
56
StereoDRNettwo views2.84
33
1.49
59
14.48
41
1.33
33
1.57
63
3.44
15
2.06
56
3.69
12
2.36
26
3.26
55
1.84
16
2.79
31
1.80
32
10.88
60
3.16
43
0.41
25
0.33
22
0.39
19
0.36
23
0.59
38
0.58
37
PDISCO_ROBtwo views2.86
34
1.53
60
14.27
40
1.98
82
2.19
86
4.52
73
1.35
31
4.65
80
3.19
61
2.09
15
2.16
37
3.53
47
2.45
46
5.32
21
2.12
27
1.30
90
0.78
82
0.72
76
0.59
63
1.24
86
1.28
81
HSM-Net_RVCpermissivetwo views2.87
35
0.57
1
17.46
62
0.92
1
0.90
18
3.70
32
1.20
27
5.64
89
2.51
36
3.22
54
2.43
46
2.54
20
1.59
22
11.41
63
1.65
10
0.34
14
0.25
13
0.26
2
0.23
8
0.30
3
0.29
5
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
ccstwo views2.91
36
1.17
40
18.05
66
1.53
56
0.91
19
3.88
45
0.80
7
3.79
17
3.00
55
1.84
4
2.13
36
2.42
17
1.25
7
13.71
87
1.58
6
0.33
11
0.27
17
0.40
22
0.35
19
0.39
9
0.34
10
MFMNet_retwo views2.91
36
1.91
69
18.25
68
3.06
90
1.89
79
3.13
4
1.67
47
3.52
6
2.97
53
3.14
52
2.87
56
2.26
11
2.41
44
2.42
5
2.00
24
1.10
84
1.05
85
0.94
85
0.90
82
1.28
87
1.46
86
GANettwo views2.93
38
1.05
28
11.29
28
1.64
64
0.96
21
3.91
47
2.52
64
3.72
14
2.41
29
2.95
47
3.00
58
3.21
41
2.03
38
12.42
79
4.72
55
0.51
42
0.41
37
0.46
32
0.39
33
0.60
41
0.50
25
PWCDC_ROBbinarytwo views2.95
39
3.26
82
14.06
36
1.59
60
1.56
61
4.19
65
0.86
10
4.30
68
4.08
77
2.06
14
7.37
77
2.75
26
2.42
45
3.59
10
2.67
36
1.34
92
0.36
30
0.39
19
0.34
16
1.13
79
0.62
41
RTSCtwo views2.95
39
2.51
79
15.49
50
1.54
58
1.30
45
3.51
23
0.86
10
3.89
28
2.48
34
3.64
61
3.12
64
3.04
39
2.30
42
5.70
28
6.67
61
0.55
50
0.35
29
0.37
13
0.39
33
0.70
55
0.60
39
PSMNet_ROBtwo views2.95
39
1.38
51
18.46
71
1.19
17
1.20
37
3.80
41
1.65
46
3.54
7
2.35
24
1.98
10
1.71
8
2.75
26
1.79
31
11.43
65
2.78
37
0.42
26
0.36
30
0.50
40
0.60
64
0.59
38
0.49
23
XQCtwo views2.96
42
2.42
78
14.85
43
1.36
41
1.32
47
3.29
10
1.53
40
4.28
63
2.78
44
3.21
53
2.46
47
3.53
47
2.92
56
6.58
30
4.67
54
0.70
69
0.41
37
0.59
58
0.56
58
0.96
72
0.83
66
SGM-Foresttwo views2.96
42
0.62
5
3.49
10
1.09
7
0.80
8
4.06
58
3.30
73
4.37
72
3.09
57
4.18
71
3.27
66
4.67
71
3.86
69
11.69
71
7.91
70
0.51
42
0.46
49
0.49
38
0.39
33
0.48
21
0.49
23
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
RYNettwo views2.97
44
0.93
18
15.09
47
1.12
11
1.38
50
3.66
30
1.42
35
4.03
41
2.29
19
2.10
16
1.70
5
3.79
51
2.16
40
13.64
86
3.02
42
0.37
19
0.41
37
0.48
35
0.34
16
0.81
64
0.75
58
G-Nettwo views3.01
45
1.37
49
18.87
74
1.26
24
1.66
68
3.69
31
1.49
38
3.82
18
2.31
21
2.68
34
2.50
48
4.38
66
2.47
47
5.67
26
3.27
45
1.27
88
0.64
74
0.64
68
0.77
75
0.69
53
0.72
55
AANet_RVCtwo views3.03
46
1.72
66
12.78
31
1.17
15
1.12
32
3.79
40
1.10
22
3.86
23
3.13
58
2.62
31
4.26
74
3.96
55
1.41
12
11.43
65
5.23
59
0.94
79
0.47
52
0.34
12
0.23
8
0.40
11
0.55
32
DRN-Testtwo views3.03
46
0.99
24
18.18
67
1.35
39
1.60
64
3.59
26
1.48
36
4.13
52
2.45
31
3.47
60
1.92
21
2.58
22
2.53
48
10.88
60
2.80
38
0.39
22
0.36
30
0.44
29
0.41
39
0.55
33
0.57
34
stereogantwo views3.09
48
0.93
18
16.43
53
1.37
43
1.66
68
3.70
32
1.67
47
3.87
25
3.15
59
3.04
48
2.70
52
5.04
75
2.79
54
9.96
49
1.78
20
0.63
58
0.49
58
0.69
71
0.47
46
0.83
66
0.68
51
Anonymous Stereotwo views3.09
48
3.14
81
16.69
57
1.52
55
1.25
41
3.33
11
2.95
69
4.02
40
2.42
30
2.35
25
2.04
28
2.57
21
1.27
8
11.88
75
2.92
40
0.49
37
0.49
58
0.57
55
0.56
58
0.70
55
0.65
47
PA-Nettwo views3.09
48
1.24
45
18.84
72
1.38
44
1.19
36
3.42
14
1.75
49
3.89
28
3.27
68
1.88
7
2.10
34
3.03
38
1.88
36
11.60
67
2.88
39
0.40
23
0.54
60
0.44
29
0.62
66
0.52
26
1.02
74
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
RGCtwo views3.10
51
1.45
57
14.90
45
1.46
47
1.79
77
3.33
11
1.39
33
4.04
43
2.26
16
2.77
39
3.26
65
4.00
57
2.71
52
11.63
69
2.62
35
0.82
75
0.46
49
0.79
80
0.79
79
0.71
57
0.78
62
PVDtwo views3.12
52
1.97
73
12.80
32
1.79
72
1.63
66
3.83
43
1.34
30
3.48
5
3.88
76
5.77
78
3.56
69
3.45
44
3.29
63
5.58
25
5.30
60
0.71
70
0.69
79
0.74
78
0.77
75
0.79
63
1.13
78
NCC-stereotwo views3.16
53
1.28
46
18.31
69
1.25
23
1.75
75
4.06
58
1.40
34
3.89
28
2.22
11
2.72
36
2.55
49
3.49
45
2.65
51
9.91
48
2.99
41
1.29
89
0.65
76
0.64
68
0.86
81
0.65
48
0.71
53
GANetREF_RVCpermissivetwo views3.20
54
1.93
70
18.84
72
1.51
54
0.80
8
4.12
63
1.82
51
4.16
56
2.91
52
1.87
6
1.74
12
2.48
18
1.84
34
13.18
83
1.74
19
0.80
73
0.57
65
0.79
80
0.74
72
1.20
85
0.86
68
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
FBW_ROBtwo views3.24
55
1.02
25
9.21
20
1.49
51
1.07
30
3.78
37
1.48
36
4.31
69
3.08
56
2.46
26
2.32
45
4.20
63
2.08
39
15.19
91
7.16
69
0.84
76
0.62
73
1.33
90
0.92
83
0.99
74
1.31
82
ADCReftwo views3.30
56
1.23
44
14.90
45
1.17
15
1.41
53
3.95
51
1.64
45
3.88
27
2.36
26
3.05
50
2.21
42
2.40
15
2.81
55
5.04
18
16.63
84
0.46
32
0.34
28
0.63
64
0.62
66
0.54
30
0.67
50
CBMVpermissivetwo views3.32
57
0.96
23
21.16
86
1.32
31
0.83
12
3.81
42
3.91
81
4.47
75
2.53
37
3.34
58
2.24
43
4.45
67
2.58
49
10.07
50
1.96
23
0.45
31
0.43
43
0.48
35
0.38
30
0.52
26
0.53
29
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
DPSNettwo views3.43
58
1.37
49
12.19
29
1.39
45
1.24
39
4.15
64
2.06
56
4.51
76
3.34
69
2.69
35
1.96
22
3.52
46
5.56
77
10.68
56
9.68
74
0.93
78
0.73
80
0.40
22
0.37
29
1.06
76
0.86
68
SAMSARAtwo views3.55
59
2.16
77
14.18
37
1.86
77
1.49
56
4.62
74
2.65
66
3.58
9
2.80
47
7.05
84
3.05
61
5.16
76
4.07
70
9.18
43
4.80
56
0.64
60
0.67
77
0.59
58
0.68
69
0.89
68
0.91
71
CBMV_ROBtwo views3.62
60
0.70
6
21.42
90
1.21
19
0.89
17
4.43
72
1.76
50
4.39
73
2.79
45
5.88
79
3.38
67
4.18
62
5.12
75
10.75
57
2.32
31
0.56
51
0.55
63
0.53
50
0.51
51
0.55
33
0.46
20
DeepPrunerFtwo views3.70
61
1.72
66
14.87
44
1.33
33
1.74
74
3.70
32
1.55
43
4.20
58
14.00
94
2.02
11
1.77
13
2.74
25
2.30
42
10.88
60
6.88
66
0.68
67
0.60
71
0.78
79
0.78
78
0.81
64
0.71
53
ADCP+two views3.74
62
1.96
72
16.90
60
1.09
7
1.55
59
4.35
69
3.65
78
3.93
34
2.98
54
2.22
20
1.66
3
3.95
54
4.83
74
3.41
8
18.87
90
0.38
21
0.41
37
0.63
64
0.53
55
0.64
47
0.81
63
ADCPNettwo views3.89
63
1.63
63
20.25
79
1.28
27
1.27
44
6.43
83
2.65
66
3.93
34
2.70
42
2.78
40
2.20
41
4.28
64
6.16
80
5.56
24
10.27
77
0.65
62
1.18
88
0.71
73
1.70
92
0.84
67
1.37
84
RTStwo views3.90
64
7.04
91
17.94
64
1.92
79
2.19
86
5.00
77
2.96
70
4.29
64
3.19
61
4.05
68
3.08
62
4.94
73
3.16
60
7.64
34
6.72
62
0.66
63
0.44
45
0.51
43
0.47
46
0.95
70
0.76
59
RTSAtwo views3.90
64
7.04
91
17.94
64
1.92
79
2.19
86
5.00
77
2.96
70
4.29
64
3.19
61
4.05
68
3.08
62
4.94
73
3.16
60
7.64
34
6.72
62
0.66
63
0.44
45
0.51
43
0.47
46
0.95
70
0.76
59
ADCMidtwo views3.96
66
3.27
83
15.48
49
1.26
24
1.40
51
4.66
75
2.28
61
4.06
46
3.54
71
3.94
66
3.02
59
3.85
52
3.67
68
4.03
12
18.65
88
0.68
67
0.61
72
1.09
88
1.18
89
1.38
89
1.08
76
MADNet+two views3.97
67
9.54
96
13.83
35
1.87
78
1.69
70
4.19
65
2.87
68
4.80
84
4.10
78
2.51
27
2.67
51
4.55
70
3.41
65
8.72
39
6.92
67
1.25
87
1.27
91
1.04
86
0.93
84
1.39
90
1.85
91
SPS-STEREOcopylefttwo views3.97
67
1.44
56
20.85
84
1.72
69
1.71
72
3.24
7
2.30
62
3.55
8
2.68
41
4.79
75
3.66
70
5.44
77
2.97
57
9.11
41
9.35
73
1.15
85
1.10
86
1.06
87
1.02
87
1.16
82
1.18
79
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
NaN_ROBtwo views4.06
69
1.13
37
16.64
55
1.12
11
1.25
41
3.50
22
3.05
72
4.29
64
3.26
66
3.64
61
2.56
50
2.97
36
4.79
73
13.33
84
16.91
85
0.37
19
0.41
37
0.38
15
0.39
33
0.48
21
0.81
63
AnyNet_C32two views4.17
70
5.75
89
12.62
30
1.60
61
1.54
57
5.51
79
4.09
82
3.90
32
3.23
64
4.22
72
3.03
60
3.10
40
3.41
65
8.01
37
18.85
89
0.53
49
0.58
67
0.53
50
0.74
72
1.19
84
0.97
73
ADCStwo views4.22
71
5.33
88
16.32
51
1.64
64
1.31
46
3.91
47
3.64
77
4.21
60
2.89
51
5.11
76
2.81
55
4.46
68
3.06
59
5.21
20
19.50
91
0.63
58
0.64
74
0.69
71
0.69
71
1.12
78
1.26
80
ADCLtwo views4.35
72
1.63
63
14.21
38
1.29
29
1.26
43
5.56
80
11.16
88
3.97
37
3.37
70
6.07
80
4.67
75
4.30
65
5.72
78
5.08
19
14.51
81
0.50
40
0.48
56
0.89
84
0.77
75
0.78
61
0.76
59
AnyNet_C01two views4.37
73
7.80
93
16.40
52
1.98
82
1.40
51
6.37
82
6.47
84
4.09
49
5.03
83
4.06
70
3.78
72
4.16
61
3.31
64
6.62
31
10.00
75
0.61
55
0.67
77
0.63
64
0.67
68
2.01
92
1.31
82
SGM_RVCbinarytwo views4.57
74
0.72
7
17.46
62
1.84
75
0.58
3
4.23
67
3.45
74
4.66
81
3.66
73
7.99
85
8.82
85
7.56
82
8.74
86
12.30
78
6.84
65
0.43
27
0.38
34
0.39
19
0.35
19
0.51
24
0.48
22
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DispFullNettwo views4.64
75
3.48
84
23.95
94
3.99
92
3.03
92
3.35
13
1.06
20
4.18
57
2.80
47
6.33
82
2.94
57
4.04
58
7.81
83
3.92
11
5.16
58
2.75
93
1.61
94
3.10
95
2.68
94
3.32
93
3.31
93
CSANtwo views4.71
76
1.39
53
20.23
78
1.19
17
0.87
14
3.65
29
3.50
76
4.57
78
3.87
75
3.89
65
7.82
80
3.54
50
6.16
80
23.47
99
6.82
64
0.52
47
0.45
48
0.49
38
0.52
52
0.63
43
0.66
49
NVStereoNet_ROBtwo views4.91
77
1.40
54
16.46
54
1.47
50
1.72
73
3.46
18
1.96
53
3.89
28
3.61
72
2.90
44
24.61
100
4.81
72
4.49
72
13.57
85
4.87
57
1.33
91
1.29
92
1.70
92
1.68
91
1.41
91
1.66
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
WCMA_ROBtwo views5.12
78
0.90
15
21.33
88
1.50
52
1.33
48
4.09
60
3.49
75
4.03
41
4.59
80
6.57
83
10.95
87
12.92
89
8.41
85
9.86
47
8.78
72
0.75
71
0.58
67
0.51
43
0.50
50
0.68
52
0.57
34
MeshStereopermissivetwo views5.35
79
1.32
47
21.34
89
1.34
36
1.15
34
4.40
71
3.90
80
5.18
85
4.55
79
10.14
92
11.82
89
11.47
87
5.84
79
13.84
88
7.03
68
0.61
55
0.58
67
0.60
61
0.58
62
0.72
59
0.59
38
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
pmcnntwo views5.45
80
1.22
43
17.40
61
1.60
61
2.41
89
3.48
21
2.12
59
4.52
77
3.26
66
3.67
63
13.15
90
22.10
95
3.46
67
11.92
77
17.16
86
0.34
14
0.21
5
0.23
1
0.19
2
0.36
7
0.28
4
MDST_ROBtwo views5.53
81
0.59
4
23.11
92
2.55
88
2.69
90
21.20
96
1.98
55
4.77
83
2.65
39
9.36
89
4.25
73
4.14
60
4.45
71
22.57
95
3.68
49
0.48
36
0.33
22
0.47
33
0.44
43
0.44
15
0.44
18
MSMD_ROBtwo views5.61
82
1.08
31
16.68
56
1.21
19
0.83
12
10.14
89
2.07
58
3.93
34
6.22
87
10.37
93
8.03
82
22.22
96
9.45
87
11.66
70
4.52
53
0.66
63
0.58
67
0.66
70
0.68
69
0.65
48
0.63
44
SANettwo views5.65
83
1.72
66
20.94
85
1.26
24
0.88
15
7.20
85
3.81
79
4.60
79
11.23
91
4.56
73
8.78
84
7.49
81
6.95
82
11.85
74
18.11
87
0.49
37
0.44
45
0.50
40
0.42
42
0.90
69
0.82
65
ELAS_RVCcopylefttwo views5.72
84
1.64
65
20.40
82
2.09
86
2.01
84
4.95
76
11.47
89
4.66
81
6.31
88
8.09
86
8.01
81
8.06
85
11.65
89
8.36
38
10.22
76
0.99
83
1.17
87
0.81
82
0.76
74
1.14
80
1.58
87
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
PWCKtwo views5.85
85
9.95
97
15.22
48
2.90
89
1.56
61
5.66
81
7.69
85
5.27
87
4.74
81
3.88
64
7.51
78
8.04
83
5.28
76
15.18
90
8.40
71
3.08
95
1.51
93
2.24
93
1.37
90
5.55
96
1.96
92
ELAScopylefttwo views5.99
86
1.62
62
20.29
80
1.96
81
2.01
84
15.09
95
8.38
87
5.23
86
5.54
85
9.02
88
8.61
83
6.74
80
9.75
88
7.48
33
11.40
78
0.96
81
1.18
88
0.72
76
0.93
84
1.30
88
1.67
89
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
Abc-Nettwo views6.58
87
2.94
80
18.36
70
1.62
63
1.62
65
9.96
88
11.53
90
20.69
97
5.05
84
3.97
67
5.61
76
14.40
92
12.47
90
6.52
29
12.11
80
0.75
71
0.54
60
0.63
64
0.61
65
1.07
77
1.09
77
FC-DCNNcopylefttwo views6.64
88
0.77
10
19.26
75
1.68
66
1.64
67
8.27
87
2.35
63
6.35
90
9.23
89
10.85
94
11.27
88
19.89
93
13.33
91
12.51
82
11.79
79
0.66
63
0.57
65
0.58
57
0.56
58
0.63
43
0.61
40
Nwc_Nettwo views6.84
89
1.95
71
19.38
77
1.46
47
1.47
54
11.20
90
14.98
92
20.44
96
3.76
74
2.72
36
3.72
71
8.25
86
14.42
93
7.90
36
21.04
94
0.89
77
0.46
49
0.50
40
0.38
30
0.98
73
0.83
66
MADNet++two views7.57
90
5.80
90
13.28
33
6.38
94
5.26
94
7.18
84
7.77
86
5.32
88
5.62
86
6.11
81
7.78
79
6.02
78
8.01
84
11.74
72
16.36
83
9.36
98
4.95
96
8.57
98
5.66
96
4.83
95
5.34
95
SGM+DAISYtwo views8.36
91
3.82
86
24.83
95
2.02
84
1.75
75
14.80
94
17.22
94
4.27
62
11.79
93
13.88
96
14.38
93
11.74
88
13.83
92
10.56
54
15.12
82
1.21
86
1.19
90
1.14
89
1.13
88
1.17
83
1.37
84
edge stereotwo views9.71
92
13.24
98
20.34
81
3.90
91
4.13
93
13.93
93
17.01
93
18.59
95
9.71
90
8.47
87
9.71
86
6.31
79
15.36
94
14.63
89
20.22
92
2.98
94
2.44
95
2.26
94
2.06
93
4.37
94
4.53
94
MANEtwo views11.39
93
1.38
51
23.16
93
4.82
93
2.79
91
22.02
97
14.51
91
25.81
98
26.28
99
10.93
95
13.17
91
19.90
94
18.03
95
15.69
92
20.41
93
0.96
81
0.74
81
1.42
91
3.24
95
0.74
60
1.71
90
SGM-ForestMtwo views11.55
94
1.16
39
21.62
91
2.14
87
0.81
10
22.25
98
19.08
95
16.50
94
11.34
92
15.94
99
14.44
94
27.12
99
22.61
100
24.76
100
28.39
99
0.50
40
0.48
56
0.47
33
0.44
43
0.55
33
0.53
29
LSMtwo views12.71
95
3.52
85
20.79
83
55.63
102
84.76
102
3.63
28
6.30
83
3.73
15
4.93
82
5.55
77
13.92
92
3.53
47
3.25
62
9.72
45
4.12
51
0.64
60
0.85
83
0.71
73
0.81
80
1.03
75
26.84
100
DGTPSM_ROBtwo views15.31
96
8.85
94
28.14
98
9.90
96
18.89
95
12.44
91
33.21
98
13.48
92
17.86
97
9.50
90
22.89
98
13.78
90
19.67
96
21.26
93
23.68
95
5.08
96
8.84
97
5.62
96
11.50
97
6.72
97
14.87
96
DPSMNet_ROBtwo views15.52
97
8.89
95
31.44
99
10.07
97
18.91
96
12.44
91
33.22
99
13.51
93
17.94
98
9.58
91
22.91
99
13.79
91
19.67
96
21.43
94
23.73
96
5.17
97
8.84
97
5.63
97
11.52
100
6.79
98
14.90
97
LE_ROBtwo views18.58
98
1.21
42
48.38
101
8.60
95
23.00
99
7.32
86
19.74
96
9.94
91
38.69
101
43.53
101
21.91
97
28.40
100
40.92
102
23.18
96
54.71
102
0.34
14
0.33
22
0.32
9
0.36
23
0.42
13
0.40
15
DPSM_ROBtwo views20.78
99
20.03
99
26.57
96
23.62
98
22.46
97
31.10
99
39.97
101
33.52
100
17.41
95
15.25
97
15.80
95
24.06
97
20.71
98
23.30
97
27.62
97
10.78
99
9.41
99
11.69
99
11.50
97
14.49
99
16.36
98
DPSMtwo views20.78
99
20.03
99
26.57
96
23.62
98
22.46
97
31.10
99
39.97
101
33.52
100
17.41
95
15.25
97
15.80
95
24.06
97
20.71
98
23.30
97
27.62
97
10.78
99
9.41
99
11.69
99
11.50
97
14.49
99
16.36
98
AVERAGE_ROBtwo views44.46
101
45.91
102
46.59
100
40.59
101
38.66
100
32.99
101
30.52
97
46.19
102
43.26
102
49.59
102
48.97
102
44.62
101
39.89
101
45.04
101
43.94
100
48.74
102
50.69
102
50.80
102
49.79
102
45.54
102
46.97
101
MEDIAN_ROBtwo views47.37
102
49.08
103
49.92
102
40.32
100
39.97
101
34.95
102
33.43
100
49.17
103
46.39
103
53.06
103
52.65
103
47.65
102
42.23
104
48.35
103
47.29
101
52.10
103
54.04
103
54.28
103
53.48
103
48.76
103
50.33
102
LSM0two views49.58
103
40.04
101
53.62
103
97.15
103
128.68
104
62.75
103
79.89
103
67.63
104
35.11
100
33.02
100
32.61
101
48.43
103
41.60
103
46.60
102
55.43
103
21.63
101
18.99
101
23.62
101
23.13
101
29.47
101
52.26
103
DPSimNet_ROBtwo views102.30
104
119.72
104
143.59
104
109.42
104
94.10
103
127.22
104
127.15
104
26.62
99
102.36
104
149.58
104
83.99
104
139.34
104
137.54
105
64.62
104
82.36
104
105.18
104
66.15
104
120.15
104
69.61
104
99.58
104
77.72
104
MSMDNettwo views1.11
4