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-Stereotwo views0.36
1
0.21
1
1.70
7
0.29
3
0.18
4
0.55
1
0.26
8
0.78
5
0.36
1
0.57
11
0.67
29
0.33
1
0.24
1
0.32
1
0.23
1
0.10
3
0.08
4
0.09
2
0.07
3
0.09
1
0.12
5
R-Stereo Traintwo views0.36
1
0.21
1
1.70
7
0.29
3
0.18
4
0.55
1
0.26
8
0.78
5
0.36
1
0.57
11
0.67
29
0.33
1
0.24
1
0.32
1
0.23
1
0.10
3
0.08
4
0.09
2
0.07
3
0.09
1
0.12
5
AdaStereotwo views0.44
3
0.38
5
1.79
14
0.34
16
0.20
7
0.57
5
0.28
13
0.92
36
0.69
13
0.52
7
0.55
5
0.61
10
0.33
4
0.48
3
0.39
5
0.17
29
0.08
4
0.17
31
0.11
13
0.18
8
0.12
5
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
HITNettwo views0.46
4
0.92
40
2.09
26
0.33
10
0.14
1
0.55
1
0.22
2
0.83
17
0.56
5
0.51
6
0.49
1
0.48
5
0.37
8
0.48
3
0.54
23
0.09
1
0.06
1
0.10
7
0.08
7
0.23
22
0.09
1
DN-CSS_ROBtwo views0.48
5
1.23
76
1.99
23
0.39
35
0.22
13
0.56
4
0.19
1
0.75
3
0.61
7
0.45
1
0.62
20
0.41
3
0.31
3
0.70
5
0.32
3
0.11
6
0.08
4
0.14
14
0.13
15
0.36
53
0.12
5
MLCVtwo views0.48
5
0.89
39
1.85
16
0.29
3
0.14
1
0.59
9
0.23
6
0.72
2
0.59
6
0.66
29
0.66
28
0.55
6
0.36
5
1.03
16
0.49
19
0.09
1
0.06
1
0.09
2
0.07
3
0.19
12
0.11
2
StereoDRNet-Refinedtwo views0.50
7
0.41
7
1.50
2
0.33
10
0.21
11
0.68
29
0.26
8
0.81
10
0.68
12
0.67
31
0.68
34
0.73
26
0.42
18
1.10
20
0.59
27
0.11
6
0.08
4
0.14
14
0.14
20
0.25
29
0.21
26
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
iResNet_ROBtwo views0.50
7
1.15
72
1.73
10
0.31
8
0.21
11
0.58
6
0.22
2
0.95
47
0.82
41
0.58
15
0.62
20
0.57
7
0.39
13
0.84
10
0.40
6
0.11
6
0.08
4
0.09
2
0.06
1
0.21
15
0.13
9
DeepPruner_ROBtwo views0.50
7
0.70
20
1.94
20
0.29
3
0.26
23
0.63
17
0.31
22
0.80
8
0.48
3
0.61
19
0.55
5
0.66
18
0.38
9
0.98
12
0.40
6
0.19
39
0.14
32
0.14
14
0.14
20
0.23
22
0.17
20
iResNettwo views0.51
10
0.76
23
2.59
56
0.41
44
0.17
3
0.64
21
0.27
12
0.82
13
0.70
16
0.57
11
0.64
27
0.43
4
0.36
5
0.70
5
0.52
21
0.10
3
0.07
3
0.09
2
0.08
7
0.21
15
0.13
9
CFNet_RVCtwo views0.51
10
0.76
23
1.79
14
0.28
2
0.30
32
0.63
17
0.28
13
0.66
1
0.55
4
0.45
1
0.57
9
0.74
27
0.41
15
1.46
30
0.42
9
0.14
20
0.10
16
0.28
65
0.13
15
0.18
8
0.14
13
DLCB_ROBtwo views0.54
12
0.55
13
1.49
1
0.36
22
0.25
17
0.65
24
0.36
28
0.80
8
0.70
16
0.72
36
0.69
37
0.83
37
0.46
22
1.53
34
0.60
28
0.13
15
0.12
27
0.14
14
0.14
20
0.14
4
0.13
9
NLCA_NET_v2_RVCtwo views0.55
13
0.92
40
2.36
42
0.38
31
0.37
50
0.60
10
0.28
13
0.82
13
0.62
9
0.62
22
0.55
5
0.72
23
0.45
21
1.00
14
0.43
12
0.15
22
0.11
19
0.14
14
0.14
20
0.15
6
0.16
16
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 views0.55
13
0.93
45
2.43
46
0.37
25
0.36
45
0.60
10
0.29
18
0.81
10
0.61
7
0.63
25
0.56
8
0.72
23
0.46
22
1.03
16
0.44
13
0.15
22
0.11
19
0.14
14
0.14
20
0.16
7
0.16
16
DISCOtwo views0.57
15
0.34
3
1.87
17
0.45
52
0.34
39
0.84
49
0.35
27
0.84
19
0.78
34
0.59
17
0.63
23
1.06
61
0.53
31
1.10
20
0.82
39
0.12
11
0.09
12
0.11
12
0.10
12
0.22
18
0.18
23
ccs_robtwo views0.57
15
0.99
55
1.97
22
0.32
9
0.20
7
0.63
17
0.22
2
0.82
13
0.72
21
0.53
9
0.61
16
0.61
10
0.38
9
1.98
50
0.44
13
0.12
11
0.09
12
0.14
14
0.14
20
0.28
39
0.16
16
NVstereo2Dtwo views0.58
17
0.40
6
2.03
25
0.35
19
0.35
41
0.69
30
0.39
29
0.91
31
0.80
37
0.47
4
0.53
2
0.67
19
0.46
22
1.99
51
0.48
17
0.19
39
0.11
19
0.17
31
0.13
15
0.25
29
0.25
39
CFNettwo views0.58
17
0.96
50
1.78
13
0.33
10
0.25
17
0.63
17
0.22
2
0.88
27
0.75
27
0.46
3
0.59
12
0.58
8
0.40
14
2.29
58
0.48
17
0.12
11
0.09
12
0.15
22
0.14
20
0.30
42
0.15
14
iResNetv2_ROBtwo views0.59
19
1.43
84
2.96
66
0.40
41
0.22
13
0.58
6
0.28
13
0.91
31
0.73
23
0.57
11
0.63
23
0.60
9
0.41
15
0.74
7
0.35
4
0.11
6
0.08
4
0.10
7
0.07
3
0.43
66
0.17
20
ETE_ROBtwo views0.60
20
0.71
21
1.55
3
0.37
25
0.25
17
0.69
30
0.51
41
1.47
80
0.70
16
0.80
43
0.69
37
0.85
40
0.49
28
1.36
29
0.58
26
0.16
26
0.13
30
0.19
39
0.16
34
0.21
15
0.23
35
LALA_ROBtwo views0.60
20
0.76
23
1.69
6
0.36
22
0.29
29
0.79
44
0.55
45
0.94
40
0.77
32
0.81
46
0.67
29
0.94
51
0.49
28
0.99
13
0.73
35
0.21
46
0.15
38
0.20
45
0.18
43
0.23
22
0.21
26
ccstwo views0.60
20
1.05
61
2.11
29
0.35
19
0.20
7
0.61
14
0.24
7
0.78
5
0.70
16
0.50
5
0.61
16
0.61
10
0.36
5
2.53
68
0.44
13
0.12
11
0.10
16
0.13
13
0.13
15
0.24
27
0.13
9
XPNet_ROBtwo views0.61
23
0.72
22
1.66
5
0.37
25
0.27
26
0.67
26
0.47
37
0.89
28
0.71
20
0.78
41
0.70
40
0.81
36
0.53
31
1.70
40
0.76
36
0.21
46
0.16
39
0.16
25
0.15
31
0.20
13
0.21
26
HSMtwo views0.62
24
0.76
23
1.72
9
0.30
7
0.29
29
0.64
21
0.28
13
0.94
40
0.62
9
0.58
15
0.57
9
1.40
72
0.64
47
2.59
70
0.42
9
0.11
6
0.08
4
0.10
7
0.08
7
0.14
4
0.11
2
StereoDRNettwo views0.64
25
0.88
38
2.35
41
0.41
44
0.45
64
0.73
37
0.47
37
0.94
40
0.74
25
0.75
39
0.54
4
0.74
27
0.47
25
1.70
40
0.72
33
0.15
22
0.14
32
0.15
22
0.14
20
0.20
13
0.19
24
DRN-Testtwo views0.64
25
0.43
11
2.44
48
0.39
35
0.44
60
0.80
45
0.41
30
1.02
61
0.78
34
0.80
43
0.63
23
0.69
21
0.56
35
1.79
46
0.69
31
0.15
22
0.12
27
0.17
31
0.16
34
0.18
8
0.17
20
NOSS_ROBtwo views0.64
25
0.65
17
1.62
4
0.46
55
0.25
17
0.67
26
0.30
20
0.93
38
0.77
32
0.59
17
0.57
9
0.77
30
0.38
9
2.78
77
0.44
13
0.23
59
0.24
65
0.28
65
0.26
68
0.27
37
0.39
67
NCCL2two views0.65
28
0.94
46
1.96
21
0.43
49
0.30
32
0.71
33
0.68
58
0.84
19
0.65
11
0.70
33
0.60
14
0.86
42
0.54
33
1.56
35
0.57
25
0.20
44
0.16
39
0.67
87
0.25
64
0.22
18
0.24
37
CVANet_RVCtwo views0.65
28
0.80
33
2.13
31
0.36
22
0.26
23
0.62
15
0.42
32
0.84
19
0.72
21
0.69
32
0.68
34
0.68
20
0.52
30
2.64
71
0.42
9
0.19
39
0.12
27
0.22
52
0.17
39
0.37
58
0.15
14
PSMNet_ROBtwo views0.66
30
1.01
60
2.33
39
0.38
31
0.31
35
0.72
35
0.55
45
0.86
22
0.69
13
0.54
10
0.60
14
0.75
29
0.47
25
2.00
52
0.82
39
0.17
29
0.13
30
0.19
39
0.19
46
0.24
27
0.27
43
TDLMtwo views0.67
31
0.92
40
1.76
11
0.37
25
0.24
16
0.66
25
0.56
47
0.82
13
0.74
25
1.16
65
0.62
20
0.61
10
0.48
27
2.70
74
0.50
20
0.20
44
0.11
19
0.35
72
0.14
20
0.32
48
0.16
16
HSM-Net_RVCpermissivetwo views0.67
31
0.34
3
2.23
36
0.26
1
0.19
6
0.67
26
0.29
18
1.26
74
0.69
13
0.72
36
0.70
40
0.78
33
0.42
18
3.77
85
0.41
8
0.13
15
0.09
12
0.10
7
0.09
10
0.12
3
0.11
2
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
AANet_RVCtwo views0.68
33
1.00
57
2.09
26
0.33
10
0.25
17
0.58
6
0.32
25
0.81
10
0.80
37
0.61
19
0.94
53
0.80
35
0.41
15
2.46
66
1.02
47
0.36
79
0.14
32
0.10
7
0.09
10
0.23
22
0.22
31
stereogantwo views0.69
34
0.51
12
2.44
48
0.39
35
0.39
52
1.01
60
0.41
30
0.94
40
0.92
52
0.73
38
0.85
50
1.15
68
0.69
48
1.57
36
0.53
22
0.19
39
0.18
47
0.20
45
0.16
34
0.36
53
0.24
37
Anonymous Stereotwo views0.70
35
1.08
65
2.47
51
0.42
47
0.36
45
0.60
10
0.78
64
0.83
17
0.75
27
0.65
28
0.63
23
0.62
15
0.38
9
2.43
64
0.78
38
0.19
39
0.18
47
0.19
39
0.20
51
0.30
42
0.23
35
RYNettwo views0.70
35
0.56
14
2.25
37
0.35
19
0.41
58
0.86
51
0.43
33
0.94
40
0.82
41
0.62
22
0.53
2
0.91
49
0.60
40
2.70
74
1.00
46
0.14
20
0.11
19
0.16
25
0.14
20
0.25
29
0.27
43
PWC_ROBbinarytwo views0.70
35
0.94
46
1.88
18
0.39
35
0.39
52
0.62
15
0.30
20
0.98
53
0.94
53
0.91
52
1.29
65
0.77
30
0.62
45
1.62
38
1.03
49
0.18
34
0.11
19
0.18
35
0.13
15
0.51
79
0.22
31
CBMVpermissivetwo views0.71
38
0.76
23
2.64
57
0.33
10
0.20
7
0.78
41
0.79
65
0.98
53
0.78
34
0.85
49
0.71
42
0.86
42
1.10
65
1.46
30
0.62
29
0.18
34
0.20
54
0.21
50
0.19
46
0.25
29
0.33
58
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
RTSCtwo views0.72
39
1.00
57
2.22
34
0.40
41
0.29
29
0.82
47
0.31
22
0.96
48
1.10
67
0.92
54
0.81
46
0.78
33
0.58
37
1.23
25
1.55
59
0.18
34
0.17
46
0.14
14
0.17
39
0.43
66
0.30
48
PWCDC_ROBbinarytwo views0.72
39
1.06
63
2.09
26
0.47
57
0.36
45
0.78
41
0.26
8
0.98
53
1.34
74
0.61
19
2.07
76
0.64
16
0.61
41
0.83
8
0.71
32
0.31
74
0.14
32
0.15
22
0.12
14
0.60
83
0.21
26
PASMtwo views0.72
39
0.68
19
1.91
19
0.47
57
0.44
60
0.60
10
0.51
41
0.76
4
0.80
37
0.77
40
0.67
29
0.87
44
0.61
41
2.40
62
0.67
30
0.32
75
0.35
77
0.36
74
0.41
80
0.45
70
0.39
67
SHDtwo views0.72
39
0.77
28
2.36
42
0.45
52
0.40
56
0.71
33
0.32
25
1.02
61
1.25
73
1.07
60
0.84
49
0.90
48
0.73
51
1.09
19
0.98
44
0.21
46
0.18
47
0.20
45
0.21
53
0.32
48
0.31
51
XQCtwo views0.73
43
0.98
54
2.39
45
0.46
55
0.35
41
0.81
46
0.46
35
1.00
58
1.00
58
0.80
43
0.69
37
0.88
46
0.74
52
1.32
27
1.20
52
0.24
61
0.16
39
0.23
57
0.20
51
0.41
65
0.33
58
GANettwo views0.75
44
0.84
35
2.01
24
0.40
41
0.25
17
0.74
38
0.66
56
0.87
24
0.73
23
0.71
35
1.07
61
0.84
38
0.59
38
2.71
76
1.15
50
0.17
29
0.16
39
0.44
81
0.16
34
0.31
46
0.20
25
PDISCO_ROBtwo views0.75
44
1.06
63
2.11
29
0.64
76
0.59
79
1.13
66
0.43
33
1.14
70
1.05
62
0.64
27
0.71
42
0.96
53
0.61
41
1.21
24
0.76
36
0.37
80
0.18
47
0.27
63
0.24
58
0.56
81
0.32
53
GANetREF_RVCpermissivetwo views0.75
44
1.15
72
2.46
50
0.47
57
0.26
23
0.83
48
0.61
52
0.97
52
0.89
48
0.62
22
0.61
16
0.72
23
0.59
38
2.42
63
0.55
24
0.32
75
0.25
69
0.36
74
0.24
58
0.39
62
0.31
51
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
CBMV_ROBtwo views0.76
47
0.42
8
2.68
59
0.33
10
0.22
13
0.74
38
0.47
37
0.96
48
0.76
30
1.04
58
0.96
55
0.88
46
0.93
58
2.06
54
1.19
51
0.22
57
0.24
65
0.27
63
0.25
64
0.25
29
0.28
45
ADCReftwo views0.80
48
0.96
50
2.29
38
0.37
25
0.35
41
0.96
55
0.51
41
0.87
24
0.75
27
0.88
50
0.83
48
0.61
10
1.36
73
1.02
15
3.27
80
0.13
15
0.11
19
0.17
31
0.19
46
0.25
29
0.21
26
DANettwo views0.80
48
0.65
17
2.21
33
1.19
87
0.48
66
0.74
38
1.25
76
0.89
28
1.53
79
0.89
51
0.71
42
0.96
53
0.56
35
1.83
47
0.96
43
0.18
34
0.16
39
0.19
39
0.16
34
0.26
35
0.29
46
FBW_ROBtwo views0.83
50
0.77
28
1.77
12
0.44
50
0.30
32
0.87
53
0.46
35
1.10
68
0.91
50
0.81
46
0.75
45
1.01
57
0.61
41
3.66
84
1.23
53
0.26
65
0.23
62
0.36
74
0.25
64
0.27
37
0.62
80
PA-Nettwo views0.84
51
1.00
57
3.49
79
0.63
75
0.40
56
0.70
32
0.52
44
0.86
22
0.84
44
0.52
7
0.59
12
0.70
22
0.54
33
2.35
59
2.23
71
0.17
29
0.27
72
0.19
39
0.28
71
0.22
18
0.36
63
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
DPSNettwo views0.86
52
0.60
16
2.34
40
0.58
72
0.36
45
1.07
64
0.75
62
1.30
75
1.05
62
0.70
33
0.68
34
0.94
51
1.29
70
2.08
55
1.82
63
0.26
65
0.23
62
0.16
25
0.17
39
0.48
72
0.40
70
DeepPrunerFtwo views0.86
52
0.86
37
2.91
65
0.45
52
0.50
67
0.64
21
0.48
40
1.14
70
2.64
88
0.63
25
0.61
16
0.65
17
0.62
45
2.21
56
1.31
55
0.23
59
0.19
51
0.30
69
0.28
71
0.36
53
0.25
39
PVDtwo views0.90
54
0.85
36
2.15
32
0.53
65
0.50
67
0.99
57
0.56
47
1.15
72
1.64
81
1.41
77
1.41
70
1.11
66
1.13
67
1.35
28
1.31
55
0.26
65
0.28
73
0.28
65
0.26
68
0.37
58
0.49
77
SAMSARAtwo views0.91
55
0.97
53
2.22
34
0.80
80
0.58
78
1.32
70
1.31
77
1.01
60
0.94
53
1.36
73
0.87
51
1.38
71
0.90
56
1.72
43
1.02
47
0.24
61
0.30
74
0.22
52
0.25
64
0.36
53
0.35
61
MFMNet_retwo views0.92
56
1.28
77
2.53
53
0.82
82
0.76
83
0.90
54
0.75
62
0.98
53
1.04
61
1.03
56
1.05
60
0.84
38
0.94
60
0.83
8
0.72
33
0.68
87
0.60
85
0.63
85
0.60
84
0.77
87
0.74
83
ADCP+two views0.94
57
1.12
68
3.14
74
0.34
16
0.39
52
1.06
62
0.79
65
0.89
28
0.80
37
0.66
29
0.67
29
0.85
40
1.55
78
0.93
11
4.55
86
0.13
15
0.11
19
0.16
25
0.15
31
0.26
35
0.25
39
ADCMidtwo views0.96
58
1.32
80
2.74
60
0.38
31
0.35
41
0.99
57
0.58
50
0.94
40
0.86
46
1.23
70
1.01
57
1.01
57
1.00
62
1.05
18
3.97
83
0.21
46
0.19
51
0.42
79
0.32
75
0.40
64
0.29
46
ADCPNettwo views0.99
59
0.99
55
3.93
84
0.37
25
0.68
81
1.36
71
0.62
53
0.87
24
0.76
30
0.78
41
0.81
46
1.08
62
1.18
68
1.14
22
2.49
75
0.21
46
0.69
88
0.21
50
0.95
90
0.29
41
0.37
65
ADCLtwo views0.99
59
0.79
31
2.43
46
0.39
35
0.32
36
1.50
73
1.90
82
0.92
36
0.97
57
1.26
72
1.27
64
1.00
56
1.42
75
1.14
22
3.31
81
0.17
29
0.14
32
0.22
52
0.21
53
0.30
42
0.22
31
AnyNet_C32two views1.00
61
1.33
81
2.53
53
0.49
61
0.54
74
1.15
68
1.14
74
0.91
31
0.89
48
1.11
62
0.95
54
0.87
44
0.93
58
1.50
32
4.07
84
0.21
46
0.20
54
0.25
62
0.24
58
0.38
60
0.33
58
SGM-Foresttwo views1.04
62
0.42
8
3.59
81
0.41
44
0.32
36
1.05
61
1.32
78
0.94
40
0.83
43
1.16
65
2.23
79
1.03
59
1.39
74
2.49
67
2.35
72
0.21
46
0.21
56
0.20
45
0.19
46
0.23
22
0.32
53
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
ADCStwo views1.05
63
1.36
83
2.98
70
0.42
47
0.36
45
0.96
55
0.81
69
1.06
67
1.12
68
1.25
71
0.97
56
1.08
62
0.90
56
1.25
26
4.86
89
0.21
46
0.21
56
0.22
52
0.23
56
0.36
53
0.38
66
SPS-STEREOcopylefttwo views1.07
64
1.13
70
2.88
63
0.65
77
0.69
82
0.99
57
0.68
58
1.05
66
0.87
47
1.36
73
1.23
63
1.46
73
0.94
60
1.87
48
1.72
61
0.66
86
0.61
86
0.58
84
0.57
82
0.65
84
0.77
86
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
WCMA_ROBtwo views1.12
65
0.83
34
2.66
58
0.58
72
0.34
39
1.17
69
0.71
61
0.93
38
0.94
53
1.74
79
2.63
83
3.09
86
1.71
79
1.70
40
1.84
64
0.26
65
0.21
56
0.20
45
0.21
53
0.28
39
0.43
73
AnyNet_C01two views1.14
66
1.49
86
4.23
85
0.54
68
0.53
73
1.51
74
1.37
79
1.04
65
1.07
64
1.17
68
1.33
69
1.15
68
1.08
64
1.61
37
2.74
77
0.25
64
0.22
59
0.28
65
0.24
58
0.48
72
0.40
70
NaN_ROBtwo views1.14
66
0.94
46
2.96
66
0.53
65
0.56
77
1.13
66
0.99
71
0.98
53
1.00
58
1.09
61
0.87
51
0.77
30
1.06
63
5.90
90
2.72
76
0.16
26
0.23
62
0.16
25
0.24
58
0.22
18
0.36
63
NVStereoNet_ROBtwo views1.16
68
0.95
49
2.54
55
0.58
72
0.55
76
0.78
41
0.63
55
1.00
58
1.08
65
1.03
56
4.07
90
1.48
74
1.22
69
2.88
80
1.27
54
0.50
82
0.46
81
0.54
83
0.62
85
0.50
77
0.63
81
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
MDST_ROBtwo views1.17
69
0.42
8
3.87
83
0.54
68
0.44
60
4.11
89
0.58
50
1.31
76
0.84
44
2.12
82
1.12
62
0.98
55
0.71
50
4.24
88
0.99
45
0.16
26
0.14
32
0.19
39
0.17
39
0.18
8
0.35
61
SANettwo views1.20
70
1.21
75
3.33
77
0.38
31
0.27
26
1.73
80
0.85
70
1.10
68
2.49
87
1.16
65
1.83
74
1.72
75
1.50
77
2.37
60
2.84
78
0.18
34
0.16
39
0.16
25
0.15
31
0.32
48
0.32
53
RTSAtwo views1.21
71
1.31
78
6.39
90
0.52
63
0.51
71
1.61
76
0.79
65
1.03
63
1.37
75
1.13
63
1.30
67
1.10
64
0.85
54
2.65
72
2.08
67
0.21
46
0.24
65
0.18
35
0.18
43
0.43
66
0.32
53
MeshStereopermissivetwo views1.21
71
1.15
72
3.01
71
0.34
16
0.28
28
1.10
65
0.69
60
1.35
77
1.01
60
2.71
85
2.41
81
2.96
85
1.45
76
2.80
78
1.40
58
0.22
57
0.22
59
0.22
52
0.19
46
0.50
77
0.26
42
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
RTStwo views1.21
71
1.31
78
6.39
90
0.52
63
0.51
71
1.61
76
0.79
65
1.03
63
1.37
75
1.13
63
1.30
67
1.10
64
0.85
54
2.65
72
2.08
67
0.21
46
0.24
65
0.18
35
0.18
43
0.43
66
0.32
53
CSANtwo views1.25
74
1.12
68
3.04
73
0.51
62
0.37
50
1.06
62
1.15
75
1.22
73
1.08
65
1.17
68
1.91
75
1.04
60
1.31
72
5.96
91
2.43
74
0.26
65
0.22
59
0.24
58
0.24
58
0.31
46
0.30
48
pmcnntwo views1.32
75
1.05
61
2.77
61
0.74
79
1.43
88
0.72
35
0.56
47
0.96
48
0.91
50
0.99
55
2.92
86
6.01
91
0.70
49
2.38
61
3.32
82
0.13
15
0.10
16
0.08
1
0.06
1
0.34
52
0.22
31
Abc-Nettwo views1.34
76
1.10
66
3.03
72
0.57
71
0.43
59
2.03
84
2.00
83
3.50
90
1.15
70
1.06
59
1.56
71
2.48
82
2.23
84
1.77
45
2.01
65
0.28
71
0.19
51
0.24
58
0.23
56
0.48
72
0.45
74
MSMD_ROBtwo views1.34
76
0.96
50
2.38
44
0.39
35
0.32
36
1.96
81
0.62
53
0.91
31
1.43
77
2.94
88
2.60
82
4.77
88
1.95
81
2.43
64
1.33
57
0.29
72
0.26
70
0.32
71
0.32
75
0.30
42
0.30
48
MADNet+two views1.36
78
1.93
89
5.95
89
0.81
81
0.59
79
1.57
75
1.09
72
1.39
79
1.22
71
0.84
48
1.01
57
1.25
70
1.12
66
3.19
82
2.36
73
0.45
81
0.47
82
0.38
78
0.34
77
0.59
82
0.57
78
ELAS_RVCcopylefttwo views1.37
79
0.78
30
3.21
75
0.55
70
0.50
67
1.49
72
2.48
87
1.48
81
1.71
83
2.13
83
2.09
77
2.07
80
2.56
86
1.90
49
2.02
66
0.35
77
0.36
78
0.37
77
0.34
77
0.49
75
0.47
75
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ELAScopylefttwo views1.39
80
0.79
31
2.86
62
0.53
65
0.50
67
2.81
87
2.20
84
1.61
84
1.50
78
2.28
84
2.35
80
1.82
76
2.13
83
1.73
44
2.22
70
0.35
77
0.37
80
0.35
72
0.35
79
0.49
75
0.47
75
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
Nwc_Nettwo views1.49
81
0.92
40
3.40
78
0.47
57
0.44
60
1.98
83
2.58
89
3.38
89
1.14
69
0.91
52
1.29
65
1.88
78
2.51
85
2.02
53
5.25
90
0.24
61
0.16
39
0.18
35
0.14
20
0.45
70
0.40
70
FC-DCNNcopylefttwo views1.49
81
0.59
15
2.88
63
0.44
50
0.39
52
1.97
82
0.66
56
1.67
85
1.70
82
2.72
86
2.89
84
4.22
87
2.97
89
2.55
69
2.19
69
0.29
72
0.26
70
0.31
70
0.30
74
0.32
48
0.39
67
PWCKtwo views1.49
81
1.90
88
3.22
76
1.09
86
0.54
74
1.71
79
1.73
81
1.49
82
1.53
79
1.44
78
1.80
73
1.82
76
1.30
71
3.34
83
1.76
62
1.07
91
0.56
84
0.97
89
0.58
83
1.38
90
0.65
82
DispFullNettwo views1.67
84
2.10
90
5.22
87
1.53
89
8.90
91
0.86
51
0.31
22
0.96
48
0.96
56
1.37
76
1.01
57
1.13
67
1.98
82
1.51
33
0.91
42
0.54
84
0.31
75
1.18
90
0.63
87
1.13
89
0.80
87
SGM+DAISYtwo views1.83
85
1.43
84
3.81
82
0.87
83
1.01
85
2.38
85
2.53
88
1.36
78
1.75
84
2.92
87
3.22
88
2.65
83
2.93
88
2.22
57
3.10
79
0.72
88
0.69
88
0.63
85
0.62
85
0.77
87
0.98
88
SGM_RVCbinarytwo views1.89
86
0.92
40
2.50
52
0.96
84
0.93
84
1.63
78
1.37
79
1.52
83
12.43
92
1.85
80
1.69
72
2.40
81
1.74
80
2.81
79
1.71
60
0.53
83
0.47
82
0.42
79
0.47
81
0.74
85
0.74
83
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
edge stereotwo views2.51
87
2.33
92
5.12
86
1.45
88
1.38
87
4.78
90
4.90
92
3.91
91
2.37
85
1.91
81
2.90
85
1.89
79
3.41
90
2.92
81
4.61
87
0.94
90
0.86
91
0.85
88
0.85
89
1.40
91
1.41
90
MADNet++two views2.79
88
2.11
91
2.97
69
2.49
90
2.28
89
2.95
88
2.31
85
2.67
86
2.47
86
3.03
89
3.13
87
2.70
84
2.92
87
4.10
86
4.48
85
2.68
92
2.24
92
3.04
92
2.44
92
2.37
92
2.34
91
MANEtwo views2.90
89
1.14
71
2.96
66
1.03
85
1.01
85
5.01
91
2.45
86
5.36
92
6.88
90
3.58
90
4.76
92
5.99
90
4.99
91
4.23
87
4.61
87
0.56
85
0.67
87
0.44
81
0.78
88
0.51
79
1.03
89
SGM-ForestMtwo views2.98
90
1.70
87
3.52
80
0.67
78
0.47
65
5.18
92
3.36
90
3.30
88
2.86
89
4.44
91
3.79
89
9.80
94
5.50
92
6.29
92
6.52
91
0.27
70
0.36
78
0.24
58
0.27
70
0.38
60
0.58
79
LSMtwo views3.31
91
1.11
67
5.29
88
9.40
94
30.65
99
0.84
49
1.12
73
0.91
31
1.22
71
1.36
73
2.19
78
0.93
50
0.83
53
1.66
39
0.90
41
0.21
46
0.31
75
0.24
58
0.28
71
0.39
62
6.33
92
LE_ROBtwo views4.43
92
1.35
82
9.38
92
2.79
91
3.93
90
2.43
86
3.46
91
2.68
87
11.19
91
11.66
96
4.73
91
5.26
89
10.46
93
4.50
89
9.49
92
0.82
89
0.69
88
1.34
91
0.98
91
0.75
86
0.76
85
DGTPSM_ROBtwo views8.89
93
5.62
93
12.25
93
5.92
92
11.32
92
8.54
93
24.18
96
6.40
93
13.49
93
5.15
92
11.46
95
6.19
92
10.92
94
7.00
93
13.72
93
4.47
93
7.65
93
3.92
93
7.45
95
4.38
93
7.87
93
DPSMNet_ROBtwo views9.48
94
5.64
94
13.01
96
8.82
93
11.82
93
8.54
93
24.20
97
6.41
94
13.65
94
5.32
93
11.47
96
6.33
93
10.92
94
7.69
94
13.89
94
7.25
96
8.19
94
5.06
94
7.78
96
5.31
94
8.26
94
DPSM_ROBtwo views12.34
95
11.21
95
12.51
94
12.91
95
12.80
94
19.84
97
28.28
98
14.38
96
14.28
95
10.58
94
11.26
93
12.03
95
11.85
96
13.15
96
15.34
96
7.11
94
8.31
95
6.50
95
7.35
93
8.84
95
8.31
95
DPSMtwo views12.34
95
11.21
95
12.51
94
12.91
95
12.80
94
19.84
97
28.28
98
14.38
96
14.28
95
10.58
94
11.26
93
12.03
95
11.85
96
13.15
96
15.34
96
7.11
94
8.31
95
6.50
95
7.35
93
8.84
95
8.31
95
DPSimNet_ROBtwo views17.97
97
20.20
97
24.98
97
17.62
97
15.92
96
21.53
99
21.00
95
13.70
95
15.61
97
23.89
98
14.69
97
22.87
97
24.58
100
12.70
95
14.56
95
18.48
98
12.91
97
20.21
98
12.05
97
17.76
97
14.17
97
MEDIAN_ROBtwo views23.96
98
27.05
99
26.39
99
22.16
98
22.13
97
16.12
95
13.97
93
20.24
98
20.05
98
25.50
99
25.31
99
22.96
98
22.63
98
24.39
98
23.24
98
28.20
99
27.23
99
28.62
99
27.75
99
27.50
99
27.76
99
LSM0two views26.58
99
22.63
98
25.88
98
30.13
100
52.21
100
39.94
100
56.74
100
28.97
100
28.72
100
21.14
97
22.26
98
24.43
99
23.88
99
26.54
99
30.87
100
14.32
97
16.55
98
13.12
97
14.71
98
17.92
98
20.60
98
AVERAGE_ROBtwo views26.93
100
30.34
100
29.27
100
25.82
99
25.34
98
19.48
96
15.17
94
23.55
99
23.00
99
28.45
100
27.87
100
26.25
100
25.86
101
26.96
100
25.55
99
31.30
100
30.07
100
31.76
100
30.81
100
30.79
100
30.93
100
MSMDNettwo views0.43
20