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
AVERAGE_ROBtwo views0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
MEDIAN_ROBtwo views0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
NVstereo2Dtwo views0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
PWCKtwo views0.01
3
0.02
6
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
PWCDC_ROBbinarytwo views0.01
3
0.01
3
0.02
7
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
PWC_ROBbinarytwo views0.01
3
0.02
6
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
HITNettwo views0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
0.01
3
LSMtwo views0.02
8
0.02
6
0.02
7
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
RYNettwo views0.02
8
0.02
6
0.02
7
0.03
10
0.03
11
0.03
13
0.02
8
0.02
8
0.02
8
0.03
12
0.02
8
0.02
8
0.03
14
0.03
10
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
LSM0two views0.02
8
0.02
6
0.02
7
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
NVStereoNet_ROBtwo views0.04
11
0.04
17
0.04
15
0.04
11
0.04
15
0.04
17
0.04
18
0.04
17
0.04
17
0.04
16
0.04
18
0.04
15
0.04
18
0.04
11
0.04
12
0.04
16
0.04
17
0.04
17
0.04
17
0.04
18
0.04
17
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 views0.08
12
0.02
6
0.02
7
1.01
57
0.03
11
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.04
11
0.09
19
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
ADCStwo views0.09
13
0.02
6
0.04
15
0.90
47
0.06
19
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.14
31
0.36
52
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
RTSAtwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
ADCPNettwo views0.10
14
0.03
14
0.04
15
1.27
70
0.03
11
0.04
17
0.03
14
0.03
14
0.03
14
0.04
16
0.03
14
0.04
15
0.03
14
0.08
17
0.04
12
0.04
16
0.03
15
0.03
14
0.03
14
0.03
15
0.03
13
SHDtwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
RTStwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
RTSCtwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
XQCtwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
SAMSARAtwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
MADNet+two views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
PVDtwo views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
MADNet++two views0.10
14
0.10
21
0.10
22
0.10
12
0.10
25
0.10
21
0.10
24
0.10
21
0.10
22
0.10
22
0.10
23
0.10
21
0.10
24
0.10
20
0.10
23
0.10
21
0.10
24
0.10
21
0.10
26
0.10
21
0.10
23
ADCLtwo views0.11
24
0.03
14
0.03
13
1.60
73
0.03
11
0.03
13
0.03
14
0.03
14
0.03
14
0.03
12
0.03
14
0.03
14
0.03
14
0.05
13
0.03
11
0.03
15
0.03
15
0.03
14
0.03
14
0.03
15
0.03
13
AnyNet_C01two views0.11
24
0.02
6
0.02
7
1.62
74
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.03
12
0.02
8
0.02
8
0.02
8
0.06
14
0.04
12
0.02
8
0.02
8
0.02
8
0.02
8
0.02
8
0.03
13
ADCReftwo views0.12
26
0.03
14
0.04
15
1.71
75
0.04
15
0.03
13
0.03
14
0.04
17
0.04
17
0.04
16
0.03
14
0.04
15
0.03
14
0.06
14
0.04
12
0.04
16
0.04
17
0.04
17
0.04
17
0.03
15
0.04
17
ELAScopylefttwo views0.13
27
0.16
32
0.11
33
0.15
24
0.09
21
0.18
35
0.11
33
0.18
35
0.11
33
0.17
34
0.11
33
0.18
35
0.11
33
0.18
34
0.11
32
0.14
33
0.08
20
0.14
33
0.08
21
0.14
33
0.09
21
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
AnyNet_C32two views0.14
28
0.04
17
0.03
13
2.22
80
0.04
15
0.03
13
0.03
14
0.03
14
0.03
14
0.03
12
0.03
14
0.04
15
0.02
8
0.07
16
0.04
12
0.02
8
0.02
8
0.03
14
0.03
14
0.02
8
0.03
13
DRN-Testtwo views0.14
28
0.13
30
0.09
21
0.14
21
0.09
21
0.15
31
0.09
21
0.14
31
0.10
22
0.14
31
0.09
21
0.14
31
0.09
21
0.13
30
0.09
19
0.12
31
0.09
23
0.12
31
0.09
22
0.12
31
0.73
62
HSMtwo views0.14
28
0.16
32
0.12
35
0.17
26
0.12
38
0.17
33
0.11
33
0.17
34
0.11
33
0.17
34
0.12
34
0.17
34
0.12
35
0.17
33
0.12
33
0.16
36
0.11
35
0.16
35
0.11
36
0.16
35
0.12
36
ELAS_RVCcopylefttwo views0.14
28
0.16
32
0.10
22
0.15
24
0.10
25
0.18
35
0.11
33
0.19
36
0.11
33
0.19
36
0.12
34
0.18
35
0.11
33
0.19
35
0.12
33
0.14
33
0.08
20
0.14
33
0.09
22
0.15
34
0.09
21
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ADCP+two views0.15
32
0.04
17
0.04
15
2.20
79
0.04
15
0.04
17
0.04
18
0.04
17
0.04
17
0.04
16
0.05
19
0.04
15
0.04
18
0.08
17
0.04
12
0.04
16
0.04
17
0.04
17
0.04
17
0.04
18
0.04
17
StereoDRNettwo views0.15
32
0.14
31
0.10
22
0.14
21
0.09
21
0.15
31
0.09
21
0.14
31
0.09
21
0.14
31
0.10
23
0.14
31
0.09
21
0.14
31
0.09
19
0.13
32
0.10
24
0.13
32
0.09
22
0.13
32
0.82
66
LALA_ROBtwo views0.15
32
0.19
37
0.12
35
0.18
27
0.11
35
0.20
39
0.12
38
0.21
39
0.12
37
0.20
39
0.12
34
0.20
38
0.12
35
0.21
38
0.13
36
0.17
38
0.10
24
0.18
38
0.11
36
0.18
37
0.11
34
SGM_RVCbinarytwo views0.15
32
0.17
35
0.11
33
0.18
27
0.11
35
0.19
37
0.11
33
0.19
36
0.12
37
0.19
36
0.12
34
0.20
38
0.12
35
0.19
35
0.12
33
0.16
36
0.11
35
0.17
37
0.10
26
0.17
36
0.10
23
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DeepPruner_ROBtwo views0.16
36
0.18
36
0.13
37
0.19
29
0.13
39
0.19
37
0.13
39
0.19
36
0.13
39
0.19
36
0.13
39
0.19
37
0.13
38
0.19
35
0.13
36
0.18
39
0.13
38
0.18
38
0.13
39
0.18
37
0.13
37
DeepPrunerFtwo views0.19
37
0.08
20
0.08
20
2.23
81
0.08
20
0.08
20
0.07
20
0.07
20
0.08
20
0.08
20
0.07
20
0.08
20
0.07
20
0.11
29
0.08
18
0.08
20
0.08
20
0.08
20
0.07
20
0.08
20
0.08
20
iResNettwo views0.20
38
0.23
38
0.18
38
0.24
30
0.18
40
0.24
40
0.20
44
0.24
40
0.18
40
0.23
40
0.17
40
0.23
40
0.18
40
0.23
39
0.18
39
0.22
41
0.16
39
0.21
40
0.16
40
0.21
40
0.16
38
iResNetv2_ROBtwo views0.20
38
0.23
38
0.18
38
0.24
30
0.20
44
0.24
40
0.18
40
0.24
40
0.18
40
0.24
42
0.18
41
0.23
40
0.19
41
0.24
40
0.18
39
0.21
40
0.16
39
0.21
40
0.16
40
0.22
41
0.16
38
iResNet_ROBtwo views0.28
40
0.32
42
0.24
43
0.32
33
0.25
47
0.32
48
0.24
46
0.32
47
0.24
47
0.33
47
0.24
47
0.35
47
0.24
44
0.33
46
0.24
46
0.31
47
0.24
45
0.32
49
0.24
44
0.32
44
0.24
45
DANettwo views0.30
41
0.30
40
0.30
47
0.30
32
0.30
51
0.30
47
0.30
51
0.30
46
0.30
51
0.30
46
0.30
51
0.30
46
0.30
48
0.30
45
0.30
50
0.30
45
0.30
49
0.30
47
0.30
48
0.30
43
0.30
49
MLCVtwo views0.31
42
0.35
43
0.27
45
0.35
34
0.28
49
0.35
50
0.27
48
0.35
49
0.28
50
0.36
50
0.27
49
0.35
47
0.27
47
0.35
48
0.27
48
0.34
48
0.27
47
0.34
50
0.27
46
0.34
45
0.27
46
DN-CSS_ROBtwo views0.31
42
0.35
43
0.28
46
0.35
34
0.28
49
0.34
49
0.27
48
0.34
48
0.25
48
0.35
48
0.27
49
0.36
50
0.26
45
0.34
47
0.28
49
0.35
49
0.28
48
0.34
50
0.28
47
0.34
45
0.27
46
StereoDRNet-Refinedtwo views0.33
44
0.39
47
0.26
44
0.39
38
0.26
48
0.40
54
0.26
47
0.40
53
0.26
49
0.39
53
0.26
48
0.39
53
0.26
45
0.39
51
0.26
47
0.37
52
0.25
46
0.37
53
0.25
45
0.37
48
0.37
51
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
ETE_ROBtwo views0.35
45
0.35
43
0.35
48
0.35
34
0.35
52
0.35
50
0.35
52
0.35
49
0.35
52
0.35
48
0.35
52
0.35
47
0.35
49
0.35
48
0.35
51
0.35
49
0.35
50
0.35
52
0.35
51
0.35
47
0.35
50
XPNet_ROBtwo views0.37
46
0.37
46
0.37
49
0.37
37
0.37
53
0.37
52
0.37
53
0.37
51
0.37
53
0.37
51
0.37
53
0.37
51
0.37
50
0.37
50
0.37
53
0.37
52
0.37
51
0.37
53
0.37
52
0.37
48
0.37
51
PASMtwo views0.39
47
3.06
74
1.36
81
1.58
71
0.09
21
0.11
30
0.11
33
0.11
30
0.11
33
0.09
21
0.09
21
0.11
30
0.09
21
0.09
19
0.09
19
0.11
30
0.11
35
0.11
30
0.09
22
0.11
30
0.11
34
AdaStereotwo views0.40
48
0.40
49
0.40
51
0.40
40
0.40
54
0.40
54
0.40
54
0.40
53
0.40
55
0.40
54
0.40
54
0.40
54
0.40
51
0.40
52
0.40
54
0.40
54
0.40
54
0.40
55
0.40
53
0.40
50
0.40
53
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
SANettwo views0.50
49
0.50
50
0.50
52
0.50
41
0.50
56
0.50
56
0.50
57
0.50
55
0.50
56
0.50
55
0.50
56
0.50
55
0.50
52
0.50
53
0.50
57
0.50
55
0.50
55
0.50
57
0.50
55
0.50
51
0.50
54
PSMNet_ROBtwo views0.54
50
0.54
51
0.54
53
0.54
42
0.53
58
0.54
57
0.54
59
0.53
56
0.54
58
0.54
56
0.54
58
0.54
56
0.53
54
0.54
54
0.54
58
0.54
57
0.53
56
0.54
58
0.54
56
0.54
52
0.54
55
CSANtwo views0.80
51
0.80
52
0.80
59
0.80
44
0.80
65
0.80
60
0.80
65
0.80
60
0.80
63
0.80
60
0.80
64
0.80
60
0.80
60
0.80
58
0.80
63
0.80
59
0.80
63
0.80
61
0.80
62
0.80
55
0.80
63
NaN_ROBtwo views0.80
51
0.80
52
0.80
59
0.80
44
0.80
65
0.80
60
0.80
65
0.80
60
0.80
63
0.80
60
0.80
64
0.80
60
0.80
60
0.80
58
0.80
63
0.80
59
0.80
63
0.80
61
0.80
62
0.80
55
0.80
63
R-Stereotwo views0.81
53
0.96
57
0.67
55
0.95
52
0.67
61
0.95
66
0.68
60
0.95
66
0.68
59
0.95
66
0.68
60
0.96
66
0.68
55
0.96
64
0.68
59
0.94
66
0.68
59
0.94
67
0.68
58
0.94
60
0.68
58
R-Stereo Traintwo views0.81
53
0.96
57
0.67
55
0.95
52
0.67
61
0.95
66
0.68
60
0.95
66
0.68
59
0.95
66
0.68
60
0.96
66
0.68
55
0.96
64
0.68
59
0.94
66
0.68
59
0.94
67
0.68
58
0.94
60
0.68
58
PDISCO_ROBtwo views0.83
55
2.80
73
3.49
89
0.14
21
0.11
35
0.17
33
0.09
21
0.15
33
0.10
22
0.15
33
0.12
34
0.16
33
0.14
39
3.30
89
0.13
36
0.14
33
2.34
88
0.16
35
0.12
38
2.62
84
0.10
23
DPSM_ROBtwo views0.92
56
0.91
54
0.91
62
0.91
48
0.92
68
0.92
63
0.92
69
0.92
63
0.91
66
0.93
64
0.91
67
0.92
62
0.90
63
0.91
60
0.92
67
0.92
62
0.91
66
0.91
64
0.92
65
0.93
58
0.93
69
DPSMtwo views0.92
56
0.91
54
0.91
62
0.91
48
0.92
68
0.92
63
0.92
69
0.92
63
0.91
66
0.93
64
0.91
67
0.92
62
0.90
63
0.91
60
0.92
67
0.92
62
0.91
66
0.91
64
0.92
65
0.93
58
0.93
69
pmcnntwo views0.92
56
0.92
56
0.92
64
0.92
50
0.92
68
0.92
63
0.92
69
0.92
63
0.92
69
0.92
63
0.92
69
0.92
62
0.92
65
0.92
62
0.92
67
0.92
62
0.92
68
0.92
66
0.92
65
0.92
57
0.92
68
DGTPSM_ROBtwo views0.93
59
1.00
59
0.92
64
0.94
51
0.96
71
0.91
62
0.92
69
0.91
62
0.96
70
0.91
62
0.96
70
0.93
65
0.92
65
0.94
63
0.92
67
0.92
62
0.92
68
0.90
63
0.92
65
0.94
60
0.91
67
HSM-Net_RVCpermissivetwo views0.95
60
1.17
65
0.78
58
1.18
65
0.71
63
1.16
80
0.75
64
1.09
72
0.80
63
1.10
73
0.75
63
1.17
80
0.74
58
1.17
78
0.80
63
1.13
76
0.73
62
1.17
78
0.77
61
1.16
71
0.72
61
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
DPSimNet_ROBtwo views0.97
61
1.18
66
0.81
61
1.10
58
0.91
67
1.02
71
0.82
67
1.04
71
0.91
66
1.03
71
0.86
66
1.28
82
0.82
62
1.03
69
0.89
66
1.17
81
0.81
65
1.02
72
0.82
64
1.08
66
0.81
65
CVANet_RVCtwo views1.00
62
1.00
59
1.00
67
1.00
54
1.00
72
1.00
68
1.00
73
1.00
68
1.00
71
1.00
68
1.00
71
1.00
68
1.00
67
1.00
66
1.00
71
1.00
68
1.00
70
1.00
69
1.00
69
1.00
63
1.00
71
GANettwo views1.00
62
1.00
59
1.00
67
1.00
54
1.00
72
1.00
68
1.00
73
1.00
68
1.00
71
1.00
68
1.00
71
1.00
68
1.00
67
1.00
66
1.00
71
1.00
68
1.00
70
1.00
69
1.00
69
1.00
63
1.00
71
TDLMtwo views1.00
62
1.00
59
1.00
67
1.00
54
1.00
72
1.00
68
1.00
73
1.00
68
1.00
71
1.00
68
1.00
71
1.00
68
1.00
67
1.00
66
1.00
71
1.00
68
1.00
70
1.00
69
1.00
69
1.00
63
1.00
71
FC-DCNNcopylefttwo views1.07
65
1.09
63
0.98
66
0.89
46
0.52
57
1.41
83
0.91
68
1.21
81
1.24
84
1.96
86
2.05
88
1.91
85
1.30
81
1.27
80
1.08
74
0.80
59
0.60
57
0.58
59
0.41
54
0.68
53
0.54
55
DISCOtwo views1.11
66
0.39
47
5.28
93
0.39
38
0.20
44
0.39
53
0.27
48
0.39
52
0.22
46
0.38
52
0.20
46
0.38
52
0.20
42
6.95
93
0.22
45
0.30
45
0.21
44
0.27
46
0.21
43
5.25
91
0.21
43
MSMD_ROBtwo views1.19
67
1.10
64
0.70
57
1.10
58
0.60
60
1.10
73
0.70
62
1.10
73
0.70
61
1.10
73
0.70
62
1.10
72
0.70
57
7.00
94
0.70
61
1.10
72
0.70
61
1.10
73
0.70
60
1.10
67
0.70
60
AANet_RVCtwo views1.26
68
0.31
41
4.99
92
5.93
91
0.20
44
0.25
42
0.20
44
0.25
42
0.20
45
0.23
40
0.19
42
0.25
42
0.20
42
0.25
41
0.20
44
5.66
93
4.76
94
0.26
45
0.30
48
0.24
42
0.27
46
CFNettwo views1.37
69
5.27
77
0.19
40
5.49
88
0.19
41
0.28
43
0.19
41
0.28
43
0.19
42
0.28
43
0.19
42
0.28
43
4.35
90
0.28
42
0.19
41
0.23
42
0.17
43
0.23
42
4.21
93
4.81
88
0.17
41
ccstwo views1.40
70
5.36
78
0.19
40
5.57
89
0.19
41
0.28
43
0.19
41
0.28
43
0.19
42
0.28
43
0.19
42
0.28
43
4.47
92
0.28
42
0.19
41
0.23
42
0.16
39
0.23
42
4.23
94
5.04
89
0.16
38
ccs_robtwo views1.40
70
5.38
79
0.19
40
5.65
90
0.19
41
0.28
43
0.19
41
0.28
43
0.19
42
0.28
43
0.19
42
0.28
43
4.45
91
0.28
42
0.19
41
0.23
42
0.16
39
0.23
42
4.19
92
5.05
90
0.17
41
GANetREF_RVCpermissivetwo views1.55
72
1.94
70
1.22
80
1.88
76
1.21
84
1.88
85
1.22
84
1.88
85
1.22
83
1.88
84
1.22
82
1.88
84
1.22
78
1.89
82
1.22
83
1.87
85
1.22
82
1.88
86
1.22
81
1.88
80
1.22
84
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
SPS-STEREOcopylefttwo views1.59
73
1.91
68
1.21
78
1.94
77
1.20
82
2.00
87
1.23
85
1.99
87
1.24
84
2.00
88
1.25
84
2.03
87
1.26
80
2.00
84
1.29
85
1.90
86
1.18
80
1.89
87
1.21
80
1.89
81
1.20
82
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
DPSMNet_ROBtwo views1.60
74
1.59
67
1.70
82
1.59
72
1.59
86
1.61
84
1.61
87
1.60
84
1.60
86
1.62
83
1.59
85
1.60
83
1.60
82
1.60
81
1.59
86
1.59
83
1.60
84
1.59
84
1.59
83
1.59
78
1.59
85
MFMNet_retwo views1.81
75
1.91
68
1.71
83
1.95
78
1.70
87
1.95
86
1.70
88
1.96
86
1.74
87
1.97
87
1.72
86
1.95
86
1.71
83
1.97
83
1.71
87
1.86
84
1.62
85
1.85
85
1.66
84
1.86
79
1.64
86
FBW_ROBtwo views2.12
76
2.46
72
1.77
84
2.49
83
1.79
88
2.38
89
1.83
89
2.46
89
1.78
88
2.48
90
1.97
87
2.40
89
1.78
84
2.42
86
1.83
88
2.31
88
1.85
86
2.38
89
1.82
85
2.35
83
1.84
87
NCCL2two views2.28
77
2.27
71
2.28
85
2.28
82
2.28
89
2.27
88
2.29
90
2.28
88
2.28
89
2.27
89
2.28
89
2.28
88
2.27
85
2.27
85
2.28
89
2.28
87
2.28
87
2.27
88
2.29
86
2.27
82
2.29
88
stereogantwo views2.33
78
24.38
85
1.18
74
1.18
65
1.18
80
1.14
77
1.18
80
1.19
78
1.19
80
1.14
79
1.18
79
1.14
77
1.18
74
1.14
75
1.18
79
1.14
78
1.14
77
1.19
80
1.19
78
1.15
69
1.19
80
RPtwo views2.33
78
25.00
86
1.13
70
1.15
61
1.15
76
1.10
73
1.15
77
1.15
75
1.15
74
1.12
77
1.15
74
1.09
71
1.13
70
1.10
71
1.15
77
1.11
75
1.09
73
1.16
75
1.12
72
1.16
71
1.17
78
NCC-stereotwo views2.36
80
25.52
88
1.17
72
1.15
61
1.17
79
1.11
76
1.11
76
1.15
75
1.16
76
1.10
73
1.16
77
1.10
72
1.15
72
1.14
75
1.13
75
1.09
71
1.11
75
1.15
74
1.16
77
1.17
74
1.17
78
RGCtwo views2.36
80
25.48
87
1.19
77
1.15
61
1.15
76
1.10
73
1.16
79
1.17
77
1.16
76
1.12
77
1.16
77
1.11
76
1.13
70
1.09
70
1.18
79
1.10
72
1.15
78
1.16
75
1.12
72
1.14
68
1.12
74
Anonymous Stereotwo views2.36
80
18.47
83
11.50
95
9.80
95
0.53
58
0.28
43
0.50
57
0.58
57
0.53
57
0.59
58
0.53
57
0.66
59
0.52
53
0.62
55
0.41
55
0.36
51
0.37
51
0.31
48
0.19
42
0.20
39
0.22
44
CF-Nettwo views2.37
83
25.71
90
1.17
72
1.13
60
1.15
76
1.15
79
1.18
80
1.19
78
1.16
76
1.10
73
1.15
74
1.10
72
1.18
74
1.12
72
1.15
77
1.10
72
1.10
74
1.17
78
1.15
76
1.16
71
1.12
74
CFNet_RVCtwo views2.38
84
8.71
81
0.38
50
9.33
94
0.42
55
0.56
58
0.42
55
0.60
58
0.38
54
0.62
59
0.42
55
0.56
57
7.29
96
0.62
55
0.42
56
0.53
56
0.37
51
0.49
56
0.34
50
8.30
94
6.87
95
edge stereotwo views2.39
85
25.69
89
1.18
74
1.16
64
1.18
80
1.14
77
1.19
82
1.19
78
1.17
79
1.14
79
1.18
79
1.14
77
1.19
76
1.13
73
1.19
81
1.15
79
1.19
81
1.19
80
1.13
74
1.19
75
1.15
76
G-Nettwo views2.42
86
26.84
91
1.14
71
1.18
65
1.14
75
1.08
72
1.15
77
1.14
74
1.15
74
1.09
72
1.15
74
1.10
72
1.15
72
1.13
73
1.14
76
1.13
76
1.11
75
1.16
75
1.19
78
1.15
69
1.16
77
Abc-Nettwo views2.53
87
28.04
92
1.18
74
1.20
68
1.20
82
1.16
80
1.20
83
1.21
81
1.20
81
1.15
81
1.21
81
1.16
79
1.20
77
1.15
77
1.20
82
1.15
79
1.16
79
1.21
82
1.14
75
1.21
76
1.21
83
DPSNettwo views3.66
88
3.60
75
3.62
90
3.63
86
3.64
93
3.65
92
3.65
93
3.66
92
3.67
92
3.67
92
3.65
93
3.67
92
3.66
88
3.66
90
3.68
93
3.68
91
3.66
91
3.67
92
3.68
90
3.67
87
3.67
92
DispFullNettwo views4.61
89
4.84
76
3.12
88
5.00
87
3.57
92
4.75
93
3.03
92
7.75
94
4.17
93
4.91
93
3.09
92
6.20
93
3.95
89
6.71
92
3.51
92
5.30
92
3.67
92
5.86
94
3.63
89
5.61
92
3.46
91
CC-Net-ROBtwo views4.74
90
40.54
94
2.83
87
2.84
84
2.86
91
2.81
91
2.90
91
2.91
91
2.90
91
2.84
91
2.89
91
2.78
90
2.83
87
2.82
88
2.90
91
2.80
89
2.80
89
2.89
91
2.85
88
2.88
86
2.88
90
NLCA_NET_v2_RVCtwo views5.04
91
50.87
95
2.78
86
2.87
85
2.71
90
2.78
90
0.47
56
2.88
90
2.86
90
1.89
85
2.67
90
2.79
91
2.79
86
2.78
87
2.83
90
2.82
90
2.82
90
2.83
90
2.77
87
2.83
85
2.76
89
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
Nwc_Nettwo views5.08
92
78.59
96
1.21
78
1.22
69
1.22
85
1.20
82
1.23
85
1.23
83
1.21
82
1.20
82
1.22
82
1.19
81
1.23
79
1.18
79
1.23
84
1.19
82
1.22
82
1.22
83
1.22
81
1.23
77
1.19
80
SGM-Foresttwo views5.21
93
5.92
80
4.08
91
6.18
92
4.16
94
6.31
94
4.34
94
6.50
93
4.33
94
6.14
94
4.21
94
6.61
94
4.55
93
6.67
91
4.48
94
5.94
94
3.94
93
5.85
93
4.03
91
5.79
93
4.17
93
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
SGM+DAISYtwo views7.06
94
9.15
82
5.38
94
8.84
93
5.18
95
8.80
95
5.31
95
8.79
95
5.28
95
8.89
95
5.20
95
8.93
95
5.33
95
8.95
95
5.36
95
8.70
95
5.21
95
8.74
95
5.20
95
8.89
95
5.15
94
PA-Nettwo views11.80
95
223.51
100
0.62
54
0.59
43
0.71
63
0.59
59
0.73
63
0.67
59
0.73
62
0.55
57
0.61
59
0.60
58
0.74
58
0.63
57
0.73
62
0.66
58
0.60
57
0.69
60
0.66
57
0.72
54
0.65
57
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
MANEtwo views19.05
96
23.00
84
15.00
96
23.00
96
15.00
96
24.00
96
16.00
96
24.00
96
16.00
96
22.00
96
15.00
96
23.00
96
16.00
97
23.00
96
15.00
96
22.00
96
15.00
96
22.00
96
15.00
96
22.00
96
15.00
96
WCMA_ROBtwo views31.10
97
35.43
93
27.12
97
39.51
97
23.10
97
38.78
97
25.30
97
37.49
97
25.39
97
37.29
97
27.02
97
38.52
97
26.48
98
37.80
97
26.44
97
36.28
97
22.65
97
33.90
97
22.30
97
37.10
97
24.18
97
MDST_ROBtwo views69.83
98
87.70
97
41.95
98
113.75
100
65.62
100
75.05
98
55.25
99
75.64
99
45.04
98
71.61
98
41.75
98
72.81
98
44.06
99
68.38
98
44.63
98
101.89
100
59.57
100
107.10
100
61.05
100
104.38
100
59.38
100
NOSS_ROBtwo views102.95
99
153.00
98
121.00
100
51.00
98
44.00
98
165.00
101
127.00
100
153.00
100
119.00
100
164.00
101
125.00
100
168.00
101
120.00
101
153.00
100
117.00
100
49.00
98
44.00
98
49.00
98
44.00
98
49.00
98
44.00
98
CBMVpermissivetwo views128.50
100
1422.70
105
53.10
99
79.50
99
51.30
99
77.30
99
49.70
98
74.00
98
48.20
99
77.20
99
48.80
99
73.90
99
48.00
100
73.80
99
48.80
99
70.40
99
45.10
99
68.90
99
46.10
99
68.20
99
45.00
99
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
MeshStereopermissivetwo views159.24
101
171.00
99
160.68
101
162.58
101
160.59
101
164.01
100
160.35
101
158.51
101
158.56
101
158.34
100
160.12
101
158.56
100
159.92
102
157.26
101
158.94
101
154.38
101
158.36
101
155.75
101
159.13
101
153.67
101
154.07
101
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
DLCB_ROBtwo views284.23
102
354.61
101
207.27
102
363.24
102
206.46
102
364.72
102
210.41
102
364.72
102
210.41
102
364.81
102
208.64
102
364.81
102
208.64
103
364.72
102
210.41
102
354.70
102
205.53
102
354.70
102
205.53
102
354.70
102
205.53
102
LE_ROBtwo views396.57
103
471.28
102
329.84
103
471.48
103
308.15
103
526.83
103
322.10
103
488.15
103
323.76
103
495.46
103
317.97
103
497.17
103
320.10
104
481.62
103
326.76
103
462.71
103
298.97
103
466.16
103
285.98
103
447.62
103
289.21
103
SGM-ForestMtwo views596.69
104
677.77
103
444.52
104
699.85
104
517.25
104
732.94
104
488.29
104
770.79
104
460.11
104
750.81
104
487.98
104
792.79
104
499.41
105
730.90
104
475.81
104
720.03
104
491.16
104
663.96
104
418.60
104
674.76
104
436.05
104
CBMV_ROBtwo views818.48
105
913.88
104
709.52
105
862.84
105
597.78
105
1073.99
105
700.52
105
1015.66
105
702.59
105
1115.65
105
760.02
105
1130.24
105
721.57
106
1037.41
105
692.65
105
814.05
105
564.29
105
843.28
105
595.31
105
915.51
105
602.92
105
MSMDNettwo views5.27
94