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 views0.57
1
0.26
1
0.63
6
0.54
7
0.36
11
0.69
2
0.54
14
1.22
2
0.72
1
1.61
37
1.50
38
0.75
2
0.53
1
0.64
1
0.44
1
0.18
2
0.15
6
0.17
2
0.15
5
0.17
1
0.17
2
R-Stereo Traintwo views0.57
1
0.26
1
0.63
6
0.54
7
0.36
11
0.69
2
0.54
14
1.22
2
0.72
1
1.61
37
1.50
38
0.75
2
0.53
1
0.64
1
0.44
1
0.18
2
0.15
6
0.17
2
0.15
5
0.17
1
0.17
2
DN-CSS_ROBtwo views0.64
3
0.32
5
1.05
27
0.55
9
0.40
20
0.54
1
0.45
3
1.21
1
0.93
3
1.03
5
1.47
35
0.69
1
0.63
4
1.76
22
0.49
3
0.18
2
0.14
3
0.25
14
0.22
15
0.22
8
0.20
8
HITNettwo views0.65
4
0.40
19
0.77
14
0.48
1
0.24
2
1.05
22
0.49
7
1.76
30
1.02
5
1.18
14
1.18
7
0.89
4
0.80
14
0.85
3
0.79
16
0.17
1
0.12
1
0.21
12
0.17
10
0.29
18
0.15
1
AdaStereotwo views0.70
5
0.43
26
0.75
12
0.73
54
0.37
14
0.92
12
0.53
12
1.85
40
1.18
10
1.12
8
1.23
9
1.07
7
0.65
6
0.96
4
0.63
7
0.31
32
0.15
6
0.32
30
0.21
13
0.29
18
0.22
13
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
CFNet_RVCtwo views0.72
6
0.32
5
0.54
2
0.66
31
0.54
37
0.83
6
0.55
19
1.37
5
1.15
7
0.93
1
1.28
13
1.57
37
0.97
25
1.35
6
0.81
20
0.26
24
0.20
20
0.33
34
0.25
19
0.30
23
0.23
15
MLCVtwo views0.72
6
0.34
10
1.10
32
0.51
2
0.23
1
0.91
10
0.39
2
1.33
4
1.23
15
1.46
30
1.55
42
1.01
6
0.64
5
1.95
27
0.80
19
0.18
2
0.12
1
0.18
5
0.14
3
0.19
4
0.17
2
DeepPruner_ROBtwo views0.74
8
0.48
38
0.87
20
0.57
12
0.44
25
0.79
5
0.54
14
1.73
25
0.96
4
1.38
26
1.27
11
1.24
16
0.73
7
1.43
7
0.73
10
0.34
42
0.28
41
0.27
16
0.26
22
0.31
26
0.31
34
iResNettwo views0.75
9
0.40
19
1.36
48
0.67
33
0.29
3
0.92
12
0.59
21
1.52
12
1.25
19
1.28
18
1.48
36
0.94
5
0.76
10
1.62
16
0.78
14
0.18
2
0.14
3
0.18
5
0.15
5
0.22
8
0.23
15
iResNet_ROBtwo views0.77
10
0.30
4
0.99
24
0.57
12
0.36
11
0.83
6
0.36
1
2.33
71
1.58
52
1.27
16
1.30
17
1.25
18
0.84
15
1.65
17
0.79
16
0.20
8
0.14
3
0.17
2
0.12
1
0.18
3
0.27
20
ccs_robtwo views0.79
11
0.32
5
0.83
17
0.60
16
0.34
8
1.03
20
0.46
4
1.81
38
1.31
26
1.12
8
1.35
23
1.18
12
0.77
12
2.63
43
0.60
6
0.23
13
0.18
16
0.27
16
0.27
26
0.23
11
0.21
9
NLCA_NET_v2_RVCtwo views0.79
11
0.43
26
1.05
27
0.64
27
0.58
46
0.95
17
0.53
12
1.53
15
1.18
10
1.37
23
1.15
4
1.43
26
0.88
17
1.61
15
0.84
21
0.25
20
0.22
27
0.27
16
0.27
26
0.29
18
0.29
26
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 views0.79
11
0.43
26
1.05
27
0.63
21
0.56
39
0.94
15
0.54
14
1.51
10
1.18
10
1.37
23
1.17
5
1.42
25
0.89
18
1.69
19
0.87
22
0.25
20
0.22
27
0.27
16
0.26
22
0.29
18
0.30
31
StereoDRNet-Refinedtwo views0.80
14
0.41
22
0.88
21
0.62
19
0.42
22
1.33
35
0.48
6
1.52
12
1.17
9
1.45
29
1.44
33
1.44
29
0.90
20
1.26
5
1.24
32
0.20
8
0.16
11
0.27
16
0.24
16
0.25
14
0.28
24
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
ccstwo views0.80
14
0.32
5
0.67
11
0.63
21
0.33
6
0.98
18
0.54
14
1.61
17
1.27
22
1.06
6
1.29
15
1.14
11
0.76
10
3.33
50
0.59
5
0.24
18
0.19
18
0.28
22
0.26
22
0.24
12
0.21
9
iResNetv2_ROBtwo views0.80
14
0.37
15
1.71
59
0.56
10
0.37
14
0.94
15
0.64
23
1.73
25
1.34
30
1.23
15
1.42
31
1.26
19
0.92
22
1.85
25
0.56
4
0.23
13
0.15
6
0.19
7
0.14
3
0.24
12
0.19
6
DLCB_ROBtwo views0.86
17
0.40
19
0.78
15
0.68
38
0.54
37
1.11
24
0.84
32
1.49
8
1.23
15
1.64
40
1.65
45
1.70
45
0.91
21
1.70
20
1.05
28
0.25
20
0.24
33
0.26
15
0.27
26
0.27
17
0.24
17
HSM-Net_RVCpermissivetwo views0.87
18
0.29
3
0.53
1
0.53
5
0.30
5
1.15
27
0.55
19
3.57
85
1.28
25
1.53
34
1.74
52
1.68
43
0.93
23
1.49
8
0.77
13
0.22
11
0.17
13
0.19
7
0.18
11
0.20
5
0.19
6
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
CBMV_ROBtwo views0.90
19
0.44
30
0.63
6
0.51
2
0.37
14
1.25
32
0.49
7
1.77
33
1.27
22
1.41
28
2.00
58
1.33
22
1.08
36
2.17
31
0.76
12
0.43
63
0.42
70
0.47
69
0.44
66
0.41
45
0.36
41
PSMNet_ROBtwo views0.90
19
0.52
42
1.17
38
0.72
48
0.61
49
1.27
33
1.12
51
1.76
30
1.12
6
1.09
7
1.28
13
1.45
30
0.89
18
1.65
17
1.42
42
0.33
37
0.25
35
0.37
48
0.39
55
0.35
32
0.29
26
NVstereo2Dtwo views0.92
21
0.41
22
1.48
53
0.72
48
0.72
55
1.20
30
0.87
35
1.75
28
1.38
34
0.99
2
1.13
1
1.26
19
0.98
28
2.21
34
0.89
23
0.43
63
0.20
20
0.38
50
0.24
16
0.57
71
0.63
76
CBMVpermissivetwo views0.95
22
0.44
30
0.66
10
0.53
5
0.35
9
1.53
42
1.63
68
1.69
21
1.42
38
2.00
56
1.60
44
1.43
26
1.17
40
1.59
13
0.99
27
0.34
42
0.34
56
0.39
55
0.33
46
0.33
31
0.32
35
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
XPNet_ROBtwo views0.96
23
0.49
40
1.09
31
0.72
48
0.57
40
1.11
24
1.14
53
1.66
20
1.23
15
1.85
49
1.39
27
1.64
41
1.11
38
1.80
24
1.42
42
0.38
54
0.30
45
0.33
34
0.29
37
0.37
35
0.34
38
ETE_ROBtwo views0.96
23
0.64
60
1.16
36
0.69
39
0.48
29
1.14
26
1.13
52
1.77
33
1.19
13
1.91
54
1.45
34
1.73
46
1.05
33
1.60
14
1.14
30
0.32
35
0.26
37
0.38
50
0.34
48
0.39
42
0.41
48
CFNettwo views0.96
23
0.36
13
1.01
25
0.62
19
0.42
22
0.99
19
0.47
5
1.90
48
1.32
28
0.99
2
1.39
27
1.10
9
0.86
16
5.58
71
0.79
16
0.23
13
0.17
13
0.28
22
0.27
26
0.26
16
0.21
9
StereoDRNettwo views0.98
26
0.47
35
1.70
58
0.76
59
0.94
74
1.51
41
0.91
37
1.92
51
1.32
28
1.61
37
1.13
1
1.47
33
0.97
25
1.75
21
1.46
45
0.28
28
0.23
30
0.29
25
0.27
26
0.35
32
0.27
20
NCCL2two views1.00
27
0.56
46
1.06
30
0.75
58
0.57
40
1.31
34
1.72
69
1.77
33
1.36
32
1.50
33
1.17
5
1.91
50
1.17
40
1.54
10
1.29
35
0.35
47
0.31
48
0.45
67
0.44
66
0.40
44
0.40
46
DRN-Testtwo views1.00
27
0.44
30
1.13
34
0.70
42
0.81
64
1.62
46
0.79
31
2.14
62
1.44
40
1.78
45
1.33
21
1.43
26
1.05
33
2.18
33
1.53
48
0.27
27
0.21
23
0.31
29
0.30
41
0.32
29
0.28
24
PA-Nettwo views1.01
29
0.57
48
1.29
43
0.64
27
0.74
57
1.19
29
0.84
32
1.69
21
1.43
39
1.16
12
1.19
8
1.32
21
1.19
43
3.42
53
1.25
33
0.29
30
0.45
74
0.33
34
0.49
75
0.32
29
0.47
60
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
LALA_ROBtwo views1.01
29
0.59
51
1.16
36
0.67
33
0.51
35
1.63
47
1.26
59
1.69
21
1.38
34
1.81
47
1.37
24
1.94
54
1.02
31
1.58
12
1.38
40
0.37
51
0.27
39
0.43
65
0.37
53
0.43
48
0.37
42
DISCOtwo views1.03
31
0.34
10
1.11
33
0.69
39
0.45
26
2.07
61
0.84
32
2.13
61
1.50
45
1.15
11
1.33
21
2.40
65
1.07
35
2.60
41
1.51
47
0.23
13
0.18
16
0.22
13
0.20
12
0.31
26
0.27
20
TDLMtwo views1.06
32
0.44
30
0.79
16
0.63
21
0.50
32
0.89
8
1.44
64
1.48
7
1.26
20
1.12
8
1.40
29
1.08
8
0.98
28
6.65
77
0.78
14
0.33
37
0.21
23
0.33
34
0.26
22
0.37
35
0.24
17
CVANet_RVCtwo views1.07
33
0.43
26
0.76
13
0.60
16
0.50
32
0.90
9
0.97
42
1.52
12
1.26
20
1.33
21
1.38
25
1.24
16
1.18
42
6.75
79
0.75
11
0.34
42
0.23
30
0.34
39
0.29
37
0.42
47
0.27
20
GANetREF_RVCpermissivetwo views1.07
33
0.86
74
1.31
44
0.82
61
0.53
36
1.41
38
1.31
61
1.76
30
1.52
47
1.27
16
1.31
19
1.46
32
1.24
45
2.03
28
0.96
26
0.56
77
0.45
74
0.67
82
0.50
76
0.88
83
0.59
73
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
DANettwo views1.08
35
0.61
58
1.32
46
0.85
64
0.72
55
1.37
37
0.51
10
1.54
16
1.15
7
1.78
45
1.54
41
2.01
57
1.11
38
2.90
48
2.00
60
0.33
37
0.30
45
0.40
60
0.30
41
0.44
52
0.45
55
NaN_ROBtwo views1.13
36
0.55
44
1.15
35
0.64
27
0.48
29
1.61
45
2.03
80
1.72
24
1.48
42
2.09
57
1.27
11
1.40
24
1.08
36
3.94
57
1.30
36
0.28
28
0.34
56
0.28
22
0.28
33
0.30
23
0.33
37
stereogantwo views1.15
37
0.42
24
1.40
49
0.86
65
0.84
66
2.38
68
0.71
28
1.88
44
1.73
58
1.72
43
1.94
57
2.31
62
1.29
49
2.07
29
1.08
29
0.36
48
0.35
60
0.39
55
0.31
44
0.51
64
0.46
58
AANet_RVCtwo views1.16
38
0.52
42
1.49
54
0.65
30
0.45
26
0.75
4
0.69
27
1.42
6
1.55
51
1.34
22
1.87
56
1.53
35
0.78
13
7.43
81
1.38
40
0.25
20
0.17
13
0.19
7
0.16
8
0.22
8
0.26
19
RYNettwo views1.17
39
0.49
40
1.24
40
0.70
42
0.97
75
1.71
49
0.92
38
1.91
49
1.46
41
1.38
26
1.14
3
1.88
48
1.30
50
4.24
60
1.87
55
0.26
24
0.21
23
0.35
42
0.27
26
0.55
69
0.54
68
HSMtwo views1.19
40
0.32
5
0.65
9
0.57
12
0.38
18
1.23
31
0.59
21
2.25
68
1.19
13
1.30
20
1.25
10
1.94
54
1.03
32
9.32
87
0.71
9
0.22
11
0.16
11
0.20
10
0.16
8
0.21
7
0.21
9
SGM-Foresttwo views1.24
41
0.38
17
0.60
3
0.52
4
0.37
14
1.56
44
1.30
60
1.65
19
1.36
32
2.11
58
1.79
53
1.63
40
0.93
23
7.48
82
0.95
25
0.36
48
0.37
63
0.39
55
0.34
48
0.37
35
0.32
35
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
PWC_ROBbinarytwo views1.24
41
0.76
71
2.46
68
0.72
48
0.78
63
1.15
27
0.49
7
1.75
28
1.92
67
2.31
61
3.07
73
1.59
38
1.40
52
2.46
38
2.10
63
0.33
37
0.22
27
0.35
42
0.24
16
0.38
40
0.41
48
NOSS_ROBtwo views1.26
43
0.46
34
0.62
5
0.56
10
0.38
18
1.06
23
0.78
29
1.73
25
1.40
36
1.01
4
1.43
32
1.10
9
0.75
8
10.62
91
0.65
8
0.45
67
0.43
71
0.48
70
0.45
68
0.44
52
0.41
48
Anonymous Stereotwo views1.28
44
0.75
69
2.45
67
0.74
56
0.77
61
0.93
14
1.77
71
1.51
10
1.31
26
1.67
41
1.29
15
1.20
14
0.75
8
6.29
74
1.68
52
0.37
51
0.36
62
0.38
50
0.40
58
0.43
48
0.45
55
RTSCtwo views1.29
45
0.73
64
2.44
66
0.83
62
0.57
40
1.85
56
0.65
24
1.88
44
2.00
69
2.20
59
1.66
46
1.60
39
1.26
47
2.33
36
3.92
73
0.32
35
0.23
30
0.27
16
0.29
37
0.37
35
0.37
42
PWCDC_ROBbinarytwo views1.29
45
0.75
69
1.33
47
1.05
79
0.88
69
1.73
50
0.52
11
1.89
46
3.28
81
1.29
19
4.88
82
1.22
15
1.46
55
1.77
23
1.45
44
0.60
79
0.26
37
0.29
25
0.25
19
0.52
65
0.37
42
SHDtwo views1.31
47
0.71
63
2.34
64
0.86
65
0.75
60
1.77
52
0.65
24
2.22
67
2.50
77
2.40
64
1.71
50
1.91
50
1.60
59
1.93
26
2.37
66
0.38
54
0.31
48
0.38
50
0.40
58
0.48
57
0.50
65
PDISCO_ROBtwo views1.32
48
0.65
61
1.44
50
1.36
85
1.46
88
2.55
71
0.92
38
2.45
74
2.07
71
1.37
23
1.49
37
1.65
42
1.00
30
2.82
46
1.35
39
1.01
85
0.27
39
0.54
76
0.47
71
0.71
82
0.73
81
PASMtwo views1.34
49
0.92
77
2.62
71
0.86
65
0.87
67
1.03
20
1.08
47
1.50
9
1.51
46
1.82
48
1.41
30
1.68
43
1.26
47
4.87
65
1.34
38
0.61
80
0.65
84
0.68
83
0.78
83
0.69
80
0.68
80
ADCReftwo views1.35
50
0.55
44
2.37
65
0.72
48
0.77
61
1.94
57
0.78
29
1.91
49
1.27
22
1.57
36
1.59
43
1.36
23
1.41
53
1.54
10
7.53
81
0.23
13
0.20
20
0.32
30
0.30
41
0.29
18
0.29
26
GANettwo views1.36
51
0.47
35
0.86
18
0.67
33
0.45
26
1.34
36
1.73
70
1.64
18
1.24
18
1.49
32
2.38
65
1.55
36
1.20
44
9.30
86
0.93
24
0.30
31
0.31
48
0.33
34
0.28
33
0.39
42
0.30
31
XQCtwo views1.37
52
0.87
76
2.56
69
0.84
63
0.69
54
1.83
55
0.89
36
2.07
60
1.88
64
1.88
50
1.53
40
1.91
50
1.55
56
2.85
47
3.05
68
0.48
70
0.29
44
0.44
66
0.39
55
0.69
80
0.60
75
MFMNet_retwo views1.40
53
1.14
84
1.68
57
1.39
86
1.07
76
1.48
40
1.43
63
1.87
42
1.82
63
2.26
60
2.19
62
1.51
34
1.86
67
1.52
9
1.25
33
0.93
84
0.83
86
0.84
85
0.80
84
0.99
86
1.05
86
CSANtwo views1.41
54
0.59
51
1.26
41
0.70
42
0.50
32
1.76
51
1.78
72
1.77
33
1.90
66
2.35
62
2.32
63
1.92
53
1.94
69
5.56
70
1.49
46
0.43
63
0.35
60
0.40
60
0.41
60
0.43
48
0.41
48
SAMSARAtwo views1.55
55
0.84
72
2.61
70
1.22
84
1.07
76
2.50
70
1.93
77
2.05
59
1.89
65
2.75
72
1.66
46
2.81
71
1.69
62
3.08
49
1.97
59
0.46
68
0.51
80
0.42
64
0.46
70
0.58
72
0.59
73
DeepPrunerFtwo views1.60
56
0.73
64
5.81
83
0.88
68
1.19
84
0.91
10
0.95
41
2.19
66
5.24
88
1.17
13
1.30
17
1.19
13
1.30
50
4.90
66
1.66
51
0.42
62
0.34
56
0.52
74
0.47
71
0.41
45
0.40
46
SPS-STEREOcopylefttwo views1.67
57
1.04
79
1.23
39
1.21
83
1.13
82
1.82
54
1.11
49
2.16
65
1.52
47
3.07
78
2.32
63
2.82
72
1.80
64
4.11
59
2.04
61
1.03
87
0.99
89
0.97
86
0.94
85
1.03
88
1.02
84
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
MDST_ROBtwo views1.67
57
0.34
10
1.58
56
0.92
70
0.58
46
4.26
87
1.11
49
2.83
78
1.41
37
4.49
83
2.55
67
1.45
30
0.97
25
7.79
83
1.17
31
0.31
32
0.25
35
0.37
48
0.33
46
0.31
26
0.30
31
SGM_RVCbinarytwo views1.68
59
0.36
13
0.61
4
0.59
15
0.29
3
2.25
66
1.18
56
2.38
72
1.68
55
3.60
79
2.67
68
3.69
77
2.42
75
8.13
84
1.90
56
0.31
32
0.31
48
0.30
28
0.28
33
0.30
23
0.29
26
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DPSNettwo views1.69
60
0.59
51
4.33
77
0.70
42
0.57
40
2.02
60
1.62
67
3.35
84
1.78
60
1.48
31
1.38
25
2.17
60
2.84
77
4.77
63
3.66
72
0.48
70
0.31
48
0.29
25
0.25
19
0.64
76
0.48
62
PVDtwo views1.70
61
0.84
72
2.26
63
0.94
73
0.83
65
2.48
69
0.97
42
2.59
76
3.35
82
3.01
76
2.77
69
2.34
63
2.47
76
2.67
45
3.14
70
0.47
69
0.45
74
0.59
81
0.48
74
0.59
74
0.75
82
FBW_ROBtwo views1.71
62
0.62
59
1.45
51
0.67
33
0.61
49
1.81
53
1.08
47
2.52
75
1.70
56
1.77
44
1.73
51
1.98
56
1.25
46
12.57
93
1.84
54
0.34
42
0.37
63
0.56
79
0.38
54
0.45
54
0.43
53
ADCLtwo views1.72
63
0.57
48
3.66
73
0.67
33
0.61
49
2.61
73
1.85
75
1.96
54
1.67
54
1.90
53
1.83
55
2.06
59
1.90
68
2.13
30
9.15
82
0.26
24
0.24
33
0.34
39
0.32
45
0.38
40
0.35
40
WCMA_ROBtwo views1.84
64
0.47
35
1.31
44
0.69
39
0.57
40
2.11
62
1.14
53
1.87
42
1.73
58
4.04
81
6.43
86
5.62
85
3.61
80
2.59
40
1.94
57
0.52
73
0.44
73
0.39
55
0.42
62
0.48
57
0.47
60
ADCPNettwo views1.85
65
0.69
62
8.07
85
0.71
47
0.58
46
2.79
76
1.22
58
1.94
53
1.54
50
1.88
50
1.69
48
2.46
67
2.04
71
2.37
37
6.48
79
0.38
54
0.37
63
0.38
50
0.43
64
0.47
56
0.45
55
SANettwo views1.88
66
0.57
48
2.18
62
0.63
21
0.40
20
2.80
77
1.57
66
2.39
73
5.88
89
2.36
63
3.57
76
3.72
78
2.97
78
4.74
62
1.76
53
0.34
42
0.30
45
0.32
30
0.29
37
0.48
57
0.37
42
NVStereoNet_ROBtwo views1.88
66
1.07
80
1.45
51
1.07
81
1.17
83
1.47
39
1.48
65
1.89
46
1.93
68
1.94
55
4.67
80
3.08
74
1.80
64
6.51
76
1.65
50
1.06
89
0.86
87
1.08
88
1.32
88
0.93
84
1.26
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
ADCMidtwo views1.89
68
0.73
64
5.06
79
0.72
48
0.67
53
2.00
59
1.14
53
1.97
56
1.48
42
2.86
74
1.81
54
2.37
64
2.05
72
2.24
35
10.39
84
0.33
37
0.31
48
0.39
55
0.43
64
0.50
61
0.44
54
AnyNet_C32two views1.96
69
1.10
82
5.17
80
0.90
69
1.11
79
2.12
63
2.37
83
1.85
40
1.48
42
2.45
65
1.70
49
2.05
58
1.64
61
2.50
39
10.28
83
0.37
51
0.34
56
0.35
42
0.41
60
0.49
60
0.54
68
ADCP+two views1.99
70
0.48
38
7.73
84
0.66
31
0.91
71
2.30
67
1.05
46
1.82
39
1.34
30
1.54
35
1.32
20
1.74
47
1.61
60
2.17
31
13.37
87
0.24
18
0.21
23
0.32
30
0.28
33
0.37
35
0.29
26
MSMD_ROBtwo views1.99
70
0.59
51
0.89
22
0.70
42
0.61
49
3.95
86
0.94
40
1.80
37
1.79
61
4.41
82
4.43
78
8.17
87
3.66
81
3.40
51
1.33
37
0.52
73
0.48
77
0.54
76
0.58
81
0.53
66
0.50
65
Abc-Nettwo views2.19
72
0.86
74
3.89
75
0.97
74
0.90
70
3.64
83
2.93
86
4.37
87
2.27
74
2.58
68
2.87
72
4.12
80
3.94
82
5.22
68
2.28
65
0.53
75
0.37
63
0.48
70
0.47
71
0.54
68
0.54
68
ADCStwo views2.25
73
0.99
78
5.55
81
0.81
60
0.74
57
2.19
65
1.80
73
2.27
69
2.14
73
2.80
73
2.13
61
2.40
65
2.03
70
2.64
44
13.69
89
0.41
59
0.39
68
0.40
60
0.45
68
0.58
72
0.56
72
MeshStereopermissivetwo views2.26
74
0.56
46
0.92
23
0.61
18
0.49
31
2.60
72
1.02
44
3.18
82
2.05
70
7.40
91
4.98
83
7.58
86
2.98
79
6.37
75
1.96
58
0.43
63
0.43
71
0.45
67
0.39
55
0.43
48
0.42
52
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
AnyNet_C01two views2.26
74
1.78
87
10.22
86
0.93
71
0.74
57
2.92
79
2.20
81
2.29
70
1.80
62
2.53
67
2.41
66
2.71
70
1.73
63
3.41
52
6.72
80
0.41
59
0.39
68
0.40
60
0.42
62
0.67
79
0.63
76
DispFullNettwo views2.34
76
2.85
89
3.66
73
3.08
92
2.27
90
1.97
58
0.67
26
2.00
57
1.70
56
2.52
66
2.00
58
2.23
61
5.86
87
3.44
55
1.58
49
1.01
85
0.32
55
2.64
91
1.94
91
2.85
90
2.24
89
pmcnntwo views2.37
77
0.42
24
2.71
72
0.63
21
0.43
24
1.54
43
1.19
57
2.01
58
1.53
49
2.64
71
7.60
88
18.07
94
1.45
54
3.44
55
2.78
67
0.18
2
0.15
6
0.15
1
0.12
1
0.20
5
0.18
5
ELAScopylefttwo views2.37
77
0.74
67
1.54
55
0.99
76
0.92
72
4.78
88
2.33
82
3.33
83
3.37
83
5.05
86
4.77
81
3.67
76
4.53
84
3.42
53
4.50
75
0.62
81
0.53
81
0.55
78
0.53
79
0.63
75
0.64
79
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SGM+DAISYtwo views2.41
79
1.07
80
2.14
61
1.18
82
1.10
78
2.14
64
1.36
62
1.96
54
1.66
53
5.46
87
7.07
87
4.95
84
4.32
83
5.50
69
2.26
64
1.04
88
1.02
90
0.99
87
0.97
86
1.02
87
1.04
85
ELAS_RVCcopylefttwo views2.48
80
0.74
67
1.90
60
1.05
79
0.92
72
2.76
74
4.01
89
2.76
77
4.11
85
4.69
84
4.63
79
4.62
82
5.96
88
3.97
58
3.96
74
0.63
82
0.53
81
0.56
79
0.52
78
0.64
76
0.63
76
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
RTSAtwo views2.67
81
1.38
85
15.28
89
0.99
76
1.26
86
3.78
84
1.96
78
2.15
63
2.62
79
2.61
69
2.85
70
2.60
68
1.59
57
6.27
72
5.71
77
0.40
57
0.28
41
0.35
42
0.35
50
0.50
61
0.48
62
RTStwo views2.67
81
1.38
85
15.28
89
0.99
76
1.26
86
3.78
84
1.96
78
2.15
63
2.62
79
2.61
69
2.85
70
2.60
68
1.59
57
6.27
72
5.71
77
0.40
57
0.28
41
0.35
42
0.35
50
0.50
61
0.48
62
MADNet+two views2.69
83
3.22
92
11.53
87
1.44
88
1.12
81
3.19
80
1.81
74
2.85
79
2.37
75
1.68
42
2.07
60
2.85
73
2.18
73
7.41
80
5.23
76
0.79
83
0.76
85
0.69
84
0.65
82
0.98
85
0.95
83
FC-DCNNcopylefttwo views2.72
84
0.59
51
0.86
18
0.73
54
0.57
40
2.91
78
1.04
45
3.01
81
2.60
78
6.83
88
7.64
89
9.92
89
6.33
89
5.21
67
3.14
70
0.53
75
0.48
77
0.51
73
0.50
76
0.53
66
0.51
67
Nwc_Nettwo views2.85
85
0.60
57
4.25
76
0.97
74
0.87
67
3.40
81
3.39
87
4.53
89
2.07
71
1.89
52
3.11
74
4.49
81
5.09
85
4.67
61
15.32
90
0.48
70
0.31
48
0.34
39
0.27
26
0.45
54
0.46
58
PWCKtwo views2.88
86
2.05
88
5.72
82
2.08
89
1.11
79
3.55
82
3.90
88
2.86
80
3.54
84
3.03
77
3.63
77
3.10
75
2.31
74
8.15
85
3.09
69
2.12
91
0.91
88
1.75
90
0.98
87
2.57
89
1.18
87
MADNet++two views5.01
87
3.14
91
4.91
78
4.55
93
3.94
91
5.20
89
4.41
90
4.40
88
4.13
86
4.95
85
6.08
84
4.74
83
5.37
86
9.40
88
10.67
85
4.65
93
4.15
92
3.97
92
3.34
92
4.09
93
4.08
91
edge stereotwo views5.63
88
2.88
90
15.35
91
2.24
90
2.24
89
11.33
91
13.87
92
7.96
92
4.75
87
3.75
80
6.08
84
4.09
79
7.76
90
4.81
64
13.55
88
1.59
90
1.56
91
1.57
89
1.53
89
2.86
91
2.74
90
MANEtwo views6.19
89
0.59
51
1.03
26
0.93
71
1.22
85
13.38
94
2.69
84
12.81
95
17.94
96
8.21
92
10.82
92
14.72
92
11.68
91
11.39
92
11.90
86
0.59
78
0.50
79
0.52
74
1.76
90
0.55
69
0.54
68
LSMtwo views6.33
90
1.10
82
13.19
88
1.41
87
68.17
99
1.64
48
1.90
76
1.92
51
2.46
76
2.99
75
3.11
74
1.88
48
1.80
64
2.61
42
2.06
62
0.41
59
0.58
83
0.50
72
0.57
80
0.64
76
17.66
97
SGM-ForestMtwo views6.52
91
0.38
17
1.26
41
0.63
21
0.35
9
13.59
95
5.32
91
6.32
91
6.70
91
11.14
93
9.18
91
22.51
97
13.93
92
16.89
95
20.07
92
0.36
48
0.37
63
0.36
47
0.35
50
0.36
34
0.34
38
LE_ROBtwo views8.05
92
0.37
15
23.53
94
0.74
56
0.33
6
2.77
75
2.87
85
3.71
86
27.24
97
25.39
97
8.87
90
8.46
88
26.71
97
6.65
77
21.99
93
0.21
10
0.19
18
0.20
10
0.21
13
0.25
14
0.22
13
DPSimNet_ROBtwo views11.20
93
4.98
93
35.43
97
2.90
91
4.52
92
7.89
90
17.09
93
4.66
90
6.06
90
14.89
96
10.93
93
17.29
93
36.81
99
13.18
94
15.68
91
2.86
92
6.29
93
5.75
95
5.09
93
3.91
92
7.70
92
DGTPSM_ROBtwo views13.01
94
8.23
94
21.87
92
9.51
94
18.10
93
11.56
92
32.01
96
10.63
93
16.20
92
7.03
89
19.48
96
10.15
90
16.66
93
9.83
89
22.42
94
4.75
94
8.68
94
5.37
93
10.25
94
6.16
94
11.36
93
DPSMNet_ROBtwo views13.02
95
8.23
94
21.91
93
9.53
95
18.11
94
11.56
92
32.02
97
10.64
94
16.23
93
7.05
90
19.49
97
10.16
91
16.66
93
9.85
90
22.43
95
4.77
95
8.68
94
5.38
94
10.26
95
6.17
95
11.36
93
DPSM_ROBtwo views19.08
96
19.14
96
25.17
95
22.15
96
20.85
95
28.17
96
37.82
98
26.56
96
16.85
94
14.14
94
14.84
94
21.70
95
18.75
95
21.87
96
26.55
96
10.16
96
9.30
96
10.47
96
10.42
96
14.03
96
12.61
95
DPSMtwo views19.08
96
19.14
96
25.17
95
22.15
96
20.85
95
28.17
96
37.82
98
26.56
96
16.85
94
14.14
94
14.84
94
21.70
95
18.75
95
21.87
96
26.55
96
10.16
96
9.30
96
10.47
96
10.42
96
14.03
96
12.61
95
AVERAGE_ROBtwo views40.43
98
42.37
99
42.73
98
36.96
99
35.58
98
30.50
98
27.61
94
36.71
99
38.39
99
44.91
99
44.36
99
37.45
99
36.70
98
41.88
98
40.35
98
45.43
99
46.81
99
47.11
99
46.35
99
42.51
99
43.87
99
MEDIAN_ROBtwo views42.00
99
44.40
100
44.29
99
35.15
98
34.95
97
31.28
99
29.60
95
35.88
98
40.02
100
47.20
100
46.64
100
37.37
98
38.48
101
44.35
100
42.45
99
48.03
100
49.39
100
49.71
100
49.03
100
45.14
100
46.65
100
LSM0two views42.82
100
38.75
98
50.81
100
47.07
100
109.38
100
57.00
100
76.02
100
53.74
100
33.71
98
29.02
98
30.41
98
43.87
100
37.22
100
43.91
99
53.54
100
20.52
98
18.62
98
21.10
98
21.11
98
28.41
98
42.22
98
MSMDNettwo views0.58
3