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
PMTNettwo 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
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
HITNettwo views0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
Vladimir Tankovich, Christian Häne, Yinda Zhang, Adarsh Kowdle, Sean Fanello, Sofien Bouaziz: HITNet: Hierarchical Iterative Tile Refinement Network for Real-time Stereo Matching. CVPR 2021
NVstereo2Dtwo views0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
PWCKtwo views0.01
4
0.02
7
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
PWCDC_ROBbinarytwo views0.01
4
0.01
4
0.02
8
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
PWC_ROBbinarytwo views0.01
4
0.02
7
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
0.01
4
LSM0two views0.02
9
0.02
7
0.02
8
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
RYNettwo views0.02
9
0.02
7
0.02
8
0.03
13
0.03
14
0.03
16
0.02
9
0.02
9
0.02
9
0.03
15
0.02
9
0.02
9
0.03
17
0.03
13
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
MFN_U_SF_DS_RVCtwo views0.02
9
0.02
7
0.02
8
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
LSMtwo views0.02
9
0.02
7
0.02
8
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
MSC_U_SF_DS_RVCtwo views0.02
9
0.02
7
0.02
8
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
NVStereoNet_ROBtwo views0.04
14
0.04
21
0.04
19
0.04
14
0.04
19
0.04
21
0.04
22
0.04
21
0.04
21
0.04
20
0.04
22
0.04
19
0.04
22
0.04
14
0.04
16
0.04
20
0.04
21
0.04
21
0.04
21
0.04
22
0.04
21
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
15
0.02
7
0.02
8
1.01
68
0.03
14
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.04
14
0.09
25
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
ADCStwo views0.09
16
0.02
7
0.04
19
0.90
58
0.06
27
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.14
41
0.36
58
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
XQCtwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
RTSAtwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
RTStwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
ADCPNettwo views0.10
17
0.03
17
0.04
19
1.27
80
0.03
14
0.04
21
0.03
17
0.03
17
0.03
17
0.04
20
0.03
17
0.04
19
0.03
17
0.08
26
0.04
16
0.04
20
0.03
18
0.03
17
0.03
17
0.03
18
0.03
16
MADNet++two views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
MADNet+two views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
RTSCtwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
PVDtwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
SHDtwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
SAMSARAtwo views0.10
17
0.10
30
0.10
31
0.10
19
0.10
35
0.10
30
0.10
34
0.10
31
0.10
32
0.10
31
0.10
33
0.10
31
0.10
34
0.10
29
0.10
29
0.10
31
0.10
34
0.10
31
0.10
36
0.10
31
0.10
33
ADCLtwo views0.11
27
0.03
17
0.03
16
1.60
84
0.03
14
0.03
16
0.03
17
0.03
17
0.03
17
0.03
15
0.03
17
0.03
17
0.03
17
0.05
20
0.03
14
0.03
18
0.03
18
0.03
17
0.03
17
0.03
18
0.03
16
BEATNet_4xtwo views0.11
27
0.03
17
0.03
16
1.60
84
0.03
14
0.03
16
0.03
17
0.03
17
0.03
17
0.03
15
0.03
17
0.03
17
0.03
17
0.05
20
0.03
14
0.03
18
0.03
18
0.03
17
0.03
17
0.03
18
0.03
16
AnyNet_C01two views0.11
27
0.02
7
0.02
8
1.62
86
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.03
15
0.02
9
0.02
9
0.02
9
0.06
22
0.04
16
0.02
9
0.02
9
0.02
9
0.02
9
0.02
9
0.03
16
FADNet-RVC-Resampletwo views0.12
30
0.04
21
0.04
19
0.04
14
0.04
19
0.04
21
0.04
22
0.04
21
0.04
21
0.04
20
0.04
22
0.04
19
0.04
22
0.04
14
1.57
94
0.04
20
0.04
21
0.04
21
0.05
27
0.04
22
0.06
27
FADNet_RVCtwo views0.12
30
0.04
21
0.04
19
0.05
18
0.04
19
0.04
21
0.04
22
0.04
21
0.04
21
0.04
20
0.04
22
0.04
19
0.05
25
0.04
14
1.65
96
0.05
26
0.04
21
0.04
21
0.04
21
0.04
22
0.04
21
FADNettwo views0.12
30
0.05
27
0.04
19
0.04
14
0.04
19
0.04
21
0.04
22
0.04
21
0.04
21
0.04
20
0.04
22
0.04
19
0.05
25
0.04
14
1.66
97
0.06
27
0.04
21
0.04
21
0.04
21
0.04
22
0.04
21
ADCReftwo views0.12
30
0.03
17
0.04
19
1.71
87
0.04
19
0.03
16
0.03
17
0.04
21
0.04
21
0.04
20
0.03
17
0.04
19
0.03
17
0.06
22
0.04
16
0.04
20
0.04
21
0.04
21
0.04
21
0.03
18
0.04
21
ELAScopylefttwo views0.13
34
0.16
42
0.11
42
0.15
32
0.09
30
0.18
45
0.11
43
0.18
45
0.11
43
0.17
44
0.11
43
0.18
45
0.11
43
0.18
44
0.11
38
0.14
43
0.08
28
0.14
43
0.08
31
0.14
43
0.09
31
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
FADNet-RVCtwo views0.13
34
0.04
21
0.04
19
0.04
14
0.04
19
0.04
21
0.05
27
0.04
21
0.04
21
0.04
20
0.05
26
0.04
19
0.05
25
0.04
14
1.71
98
0.04
20
0.04
21
0.04
21
0.04
21
0.07
27
0.04
21
HSMtwo views0.14
36
0.16
42
0.12
44
0.17
34
0.12
48
0.17
43
0.11
43
0.17
44
0.11
43
0.17
44
0.12
44
0.17
44
0.12
45
0.17
43
0.12
39
0.16
46
0.11
45
0.16
45
0.11
46
0.16
45
0.12
46
DRN-Testtwo views0.14
36
0.13
40
0.09
30
0.14
29
0.09
30
0.15
41
0.09
30
0.14
41
0.10
32
0.14
41
0.09
30
0.14
41
0.09
31
0.13
40
0.09
25
0.12
41
0.09
33
0.12
41
0.09
32
0.12
41
0.73
72
AnyNet_C32two views0.14
36
0.04
21
0.03
16
2.22
92
0.04
19
0.03
16
0.03
17
0.03
17
0.03
17
0.03
15
0.03
17
0.04
19
0.02
9
0.07
24
0.04
16
0.02
9
0.02
9
0.03
17
0.03
17
0.02
9
0.03
16
ELAS_RVCcopylefttwo views0.14
36
0.16
42
0.10
31
0.15
32
0.10
35
0.18
45
0.11
43
0.19
46
0.11
43
0.19
46
0.12
44
0.18
45
0.11
43
0.19
45
0.12
39
0.14
43
0.08
28
0.14
43
0.09
32
0.15
44
0.09
31
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
StereoDRNettwo views0.15
40
0.14
41
0.10
31
0.14
29
0.09
30
0.15
41
0.09
30
0.14
41
0.09
31
0.14
41
0.10
33
0.14
41
0.09
31
0.14
41
0.09
25
0.13
42
0.10
34
0.13
42
0.09
32
0.13
42
0.82
78
LALA_ROBtwo views0.15
40
0.19
47
0.12
44
0.18
35
0.11
45
0.20
49
0.12
48
0.21
49
0.12
47
0.20
49
0.12
44
0.20
48
0.12
45
0.21
48
0.13
42
0.17
48
0.10
34
0.18
48
0.11
46
0.18
47
0.11
44
ADCP+two views0.15
40
0.04
21
0.04
19
2.20
91
0.04
19
0.04
21
0.04
22
0.04
21
0.04
21
0.04
20
0.05
26
0.04
19
0.04
22
0.08
26
0.04
16
0.04
20
0.04
21
0.04
21
0.04
21
0.04
22
0.04
21
SGM_RVCbinarytwo views0.15
40
0.17
45
0.11
42
0.18
35
0.11
45
0.19
47
0.11
43
0.19
46
0.12
47
0.19
46
0.12
44
0.20
48
0.12
45
0.19
45
0.12
39
0.16
46
0.11
45
0.17
47
0.10
36
0.17
46
0.10
33
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DeepPruner_ROBtwo views0.16
44
0.18
46
0.13
46
0.19
37
0.13
49
0.19
47
0.13
49
0.19
46
0.13
49
0.19
46
0.13
49
0.19
47
0.13
48
0.19
45
0.13
42
0.18
49
0.13
48
0.18
48
0.13
49
0.18
47
0.13
47
DeepPrunerFtwo views0.19
45
0.08
28
0.08
28
2.23
93
0.08
28
0.08
28
0.07
28
0.07
28
0.08
29
0.08
28
0.07
28
0.08
28
0.07
28
0.11
38
0.08
23
0.08
29
0.08
28
0.08
29
0.07
28
0.08
28
0.08
29
BEATNet-Init1two views0.19
45
0.08
28
0.08
28
2.23
93
0.08
28
0.08
28
0.07
28
0.07
28
0.08
29
0.08
28
0.07
28
0.08
28
0.07
28
0.11
38
0.08
23
0.08
29
0.08
28
0.08
29
0.07
28
0.08
28
0.08
29
iResNettwo views0.20
47
0.23
48
0.18
47
0.24
38
0.18
50
0.24
50
0.20
53
0.24
50
0.18
50
0.23
50
0.17
50
0.23
50
0.18
50
0.23
49
0.18
45
0.22
51
0.16
49
0.21
50
0.16
50
0.21
50
0.16
48
iResNetv2_ROBtwo views0.20
47
0.23
48
0.18
47
0.24
38
0.20
53
0.24
50
0.18
50
0.24
50
0.18
50
0.24
52
0.18
51
0.23
50
0.19
51
0.24
50
0.18
45
0.21
50
0.16
49
0.21
50
0.16
50
0.22
51
0.16
48
SuperBtwo views0.21
49
0.10
30
2.51
99
0.12
28
0.09
30
0.10
30
0.09
30
0.08
30
0.07
28
0.10
31
0.09
30
0.09
30
0.07
28
0.07
24
0.07
22
0.07
28
0.08
28
0.07
28
0.07
28
0.08
28
0.07
28
iResNet_ROBtwo views0.28
50
0.32
52
0.24
51
0.32
41
0.25
56
0.32
57
0.24
55
0.32
56
0.24
56
0.33
56
0.24
56
0.35
56
0.24
54
0.33
55
0.24
52
0.31
56
0.24
54
0.32
58
0.24
54
0.32
54
0.24
54
DANettwo views0.30
51
0.30
50
0.30
55
0.30
40
0.30
61
0.30
56
0.30
60
0.30
55
0.30
60
0.30
55
0.30
60
0.30
55
0.30
58
0.30
54
0.30
56
0.30
54
0.30
58
0.30
56
0.30
59
0.30
53
0.30
58
DN-CSS_ROBtwo views0.31
52
0.35
53
0.28
54
0.35
42
0.28
59
0.34
58
0.27
57
0.34
58
0.25
57
0.35
57
0.27
58
0.36
59
0.26
55
0.34
56
0.28
55
0.35
58
0.28
57
0.34
59
0.28
58
0.34
55
0.27
55
MLCVtwo views0.31
52
0.35
53
0.27
53
0.35
42
0.28
59
0.35
59
0.27
57
0.35
59
0.28
59
0.36
59
0.27
58
0.35
56
0.27
57
0.35
57
0.27
54
0.34
57
0.27
56
0.34
59
0.27
56
0.34
55
0.27
55
StereoDRNet-Refinedtwo views0.33
54
0.39
57
0.26
52
0.39
46
0.26
57
0.40
63
0.26
56
0.40
63
0.26
58
0.39
62
0.26
57
0.39
62
0.26
55
0.39
60
0.26
53
0.37
61
0.25
55
0.37
63
0.25
55
0.37
58
0.37
60
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
ETE_ROBtwo views0.35
55
0.35
53
0.35
56
0.35
42
0.35
62
0.35
59
0.35
62
0.35
59
0.35
61
0.35
57
0.35
61
0.35
56
0.35
59
0.35
57
0.35
57
0.35
58
0.35
59
0.35
61
0.35
62
0.35
57
0.35
59
XPNet_ROBtwo views0.37
56
0.37
56
0.37
57
0.37
45
0.37
63
0.37
61
0.37
63
0.37
61
0.37
62
0.37
60
0.37
62
0.37
60
0.37
60
0.37
59
0.37
59
0.37
61
0.37
60
0.37
63
0.37
63
0.37
58
0.37
60
PASMtwo views0.39
57
3.06
87
1.36
92
1.58
82
0.09
30
0.11
40
0.11
43
0.11
40
0.11
43
0.09
30
0.09
30
0.11
40
0.09
31
0.09
28
0.09
25
0.11
40
0.11
45
0.11
40
0.09
32
0.11
40
0.11
44
AdaStereotwo views0.40
58
0.40
59
0.40
59
0.40
48
0.40
64
0.40
63
0.40
64
0.40
63
0.40
64
0.40
63
0.40
63
0.40
63
0.40
61
0.40
61
0.40
60
0.40
63
0.40
64
0.40
65
0.40
64
0.40
60
0.40
62
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021.
SANettwo views0.50
59
0.50
60
0.50
61
0.50
49
0.50
66
0.50
66
0.50
67
0.50
65
0.50
66
0.50
65
0.50
66
0.50
64
0.50
62
0.50
63
0.50
63
0.50
64
0.50
65
0.50
67
0.50
66
0.50
61
0.50
63
PSMNet_ROBtwo views0.54
60
0.54
61
0.54
62
0.54
50
0.53
68
0.54
67
0.54
69
0.53
66
0.54
68
0.54
66
0.54
68
0.54
65
0.53
64
0.54
64
0.54
64
0.54
67
0.53
66
0.54
68
0.54
67
0.54
62
0.54
64
G-Nettwo views0.79
61
0.79
62
0.79
69
0.79
53
0.79
76
0.79
71
0.79
76
0.79
71
0.79
74
0.79
71
0.79
75
0.79
71
0.79
71
0.79
69
0.79
70
0.79
70
0.79
74
0.79
72
0.79
74
0.79
66
0.79
73
STTStereo_v2two views0.79
61
0.79
62
0.79
69
0.79
53
0.79
76
0.79
71
0.79
76
0.79
71
0.79
74
0.79
71
0.79
75
0.79
71
0.79
71
0.79
69
0.79
70
0.79
70
0.79
74
0.79
72
0.79
74
0.79
66
0.79
73
NaN_ROBtwo views0.80
63
0.80
64
0.80
71
0.80
55
0.80
78
0.80
73
0.80
78
0.80
73
0.80
76
0.80
73
0.80
77
0.80
73
0.80
73
0.80
71
0.80
72
0.80
72
0.80
76
0.80
74
0.80
76
0.80
68
0.80
75
CSANtwo views0.80
63
0.80
64
0.80
71
0.80
55
0.80
78
0.80
73
0.80
78
0.80
73
0.80
76
0.80
73
0.80
77
0.80
73
0.80
73
0.80
71
0.80
72
0.80
72
0.80
76
0.80
74
0.80
76
0.80
68
0.80
75
R-Stereo Traintwo views0.81
65
0.96
69
0.67
64
0.95
63
0.67
71
0.95
79
0.68
70
0.95
79
0.68
69
0.95
79
0.68
70
0.96
79
0.68
65
0.96
77
0.68
65
0.94
79
0.68
69
0.94
80
0.68
69
0.94
73
0.68
67
R-Stereotwo views0.81
65
0.96
69
0.67
64
0.95
63
0.67
71
0.95
79
0.68
70
0.95
79
0.68
69
0.95
79
0.68
70
0.96
79
0.68
65
0.96
77
0.68
65
0.94
79
0.68
69
0.94
80
0.68
69
0.94
73
0.68
67
PDISCO_ROBtwo views0.83
67
2.80
86
3.49
103
0.14
29
0.11
45
0.17
43
0.09
30
0.15
43
0.10
32
0.15
43
0.12
44
0.16
43
0.14
49
3.30
103
0.13
42
0.14
43
2.34
102
0.16
45
0.12
48
2.62
98
0.10
33
pmcnntwo views0.92
68
0.92
68
0.92
76
0.92
61
0.92
81
0.92
76
0.92
82
0.92
76
0.92
82
0.92
76
0.92
82
0.92
75
0.92
78
0.92
75
0.92
76
0.92
75
0.92
81
0.92
79
0.92
79
0.92
70
0.92
80
DPSM_ROBtwo views0.92
68
0.91
66
0.91
74
0.91
59
0.92
81
0.92
76
0.92
82
0.92
76
0.91
79
0.93
77
0.91
80
0.92
75
0.90
76
0.91
73
0.92
76
0.92
75
0.91
79
0.91
77
0.92
79
0.93
71
0.93
81
DPSMtwo views0.92
68
0.91
66
0.91
74
0.91
59
0.92
81
0.92
76
0.92
82
0.92
76
0.91
79
0.93
77
0.91
80
0.92
75
0.90
76
0.91
73
0.92
76
0.92
75
0.91
79
0.91
77
0.92
79
0.93
71
0.93
81
DGTPSM_ROBtwo views0.93
71
1.00
71
0.92
76
0.94
62
0.96
84
0.91
75
0.92
82
0.91
75
0.96
83
0.91
75
0.96
83
0.93
78
0.92
78
0.94
76
0.92
76
0.92
75
0.92
81
0.90
76
0.92
79
0.94
73
0.91
79
HSM-Net_RVCpermissivetwo views0.95
72
1.17
77
0.78
68
1.18
78
0.71
74
1.16
94
0.75
75
1.09
85
0.80
76
1.10
86
0.75
74
1.17
93
0.74
69
1.17
91
0.80
72
1.13
92
0.73
73
1.17
92
0.77
73
1.16
84
0.72
71
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
DPSimNet_ROBtwo views0.97
73
1.18
78
0.81
73
1.10
70
0.91
80
1.02
84
0.82
80
1.04
84
0.91
79
1.03
84
0.86
79
1.28
94
0.82
75
1.03
82
0.89
75
1.17
94
0.81
78
1.02
85
0.82
78
1.08
79
0.81
77
GANettwo views1.00
74
1.00
71
1.00
79
1.00
65
1.00
85
1.00
81
1.00
86
1.00
81
1.00
84
1.00
81
1.00
84
1.00
81
1.00
80
1.00
79
1.00
80
1.00
81
1.00
83
1.00
82
1.00
83
1.00
76
1.00
83
CVANet_RVCtwo views1.00
74
1.00
71
1.00
79
1.00
65
1.00
85
1.00
81
1.00
86
1.00
81
1.00
84
1.00
81
1.00
84
1.00
81
1.00
80
1.00
79
1.00
80
1.00
81
1.00
83
1.00
82
1.00
83
1.00
76
1.00
83
TDLMtwo views1.00
74
1.00
71
1.00
79
1.00
65
1.00
85
1.00
81
1.00
86
1.00
81
1.00
84
1.00
81
1.00
84
1.00
81
1.00
80
1.00
79
1.00
80
1.00
81
1.00
83
1.00
82
1.00
83
1.00
76
1.00
83
DMCAtwo views1.06
77
8.20
94
0.68
66
0.68
52
0.68
73
0.68
70
0.68
70
0.69
70
0.68
69
0.68
70
0.68
70
0.68
70
0.68
65
0.68
68
0.69
67
0.68
69
0.68
69
0.68
70
0.70
71
0.68
63
0.68
67
FC-DCNNcopylefttwo views1.07
78
1.09
75
0.98
78
0.89
57
0.52
67
1.41
95
0.91
81
1.21
95
1.24
96
1.96
99
2.05
101
1.91
98
1.30
93
1.27
92
1.08
83
0.80
72
0.60
67
0.58
69
0.41
65
0.68
63
0.54
64
DISCOtwo views1.11
79
0.39
57
5.28
107
0.39
46
0.20
53
0.39
62
0.27
57
0.39
62
0.22
55
0.38
61
0.20
55
0.38
61
0.20
52
6.95
107
0.22
50
0.30
54
0.21
53
0.27
55
0.21
53
5.25
104
0.21
52
MSMD_ROBtwo views1.19
80
1.10
76
0.70
67
1.10
70
0.60
70
1.10
87
0.70
73
1.10
86
0.70
72
1.10
86
0.70
73
1.10
87
0.70
68
7.00
108
0.70
68
1.10
86
0.70
72
1.10
86
0.70
71
1.10
80
0.70
70
AANet_RVCtwo views1.26
81
0.31
51
4.99
106
5.93
104
0.20
53
0.25
52
0.20
53
0.25
52
0.20
54
0.23
50
0.19
52
0.25
52
0.20
52
0.25
51
0.20
49
5.66
107
4.76
108
0.26
54
0.30
59
0.24
52
0.27
55
CFNettwo views1.37
82
5.27
90
0.19
49
5.49
102
0.19
51
0.28
53
0.19
51
0.28
53
0.19
52
0.28
53
0.19
52
0.28
53
4.35
104
0.28
52
0.19
47
0.23
52
0.17
52
0.23
52
4.21
108
4.81
102
0.17
50
ccs_robtwo views1.40
83
5.38
91
0.19
49
5.65
103
0.19
51
0.28
53
0.19
51
0.28
53
0.19
52
0.28
53
0.19
52
0.28
53
4.45
105
0.28
52
0.19
47
0.23
52
0.16
49
0.23
52
4.19
107
5.05
103
0.17
50
RASNettwo views1.49
84
1.65
80
1.45
93
1.38
81
1.43
98
1.47
96
1.36
99
1.38
96
1.36
98
1.39
95
1.60
98
1.45
95
1.45
94
1.51
93
2.21
101
1.53
95
1.36
96
1.36
96
1.36
96
1.66
91
1.36
96
GANetREF_RVCpermissivetwo views1.55
85
1.94
83
1.22
91
1.88
88
1.21
97
1.88
98
1.22
97
1.88
98
1.22
95
1.88
97
1.22
95
1.88
97
1.22
91
1.89
95
1.22
92
1.87
98
1.22
95
1.88
99
1.22
95
1.88
93
1.22
95
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
86
1.91
81
1.21
90
1.94
89
1.20
96
2.00
100
1.23
98
1.99
100
1.24
96
2.00
101
1.25
96
2.03
100
1.26
92
2.00
97
1.29
93
1.90
99
1.18
94
1.89
100
1.21
94
1.89
94
1.20
94
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
DPSMNet_ROBtwo views1.60
87
1.59
79
1.70
94
1.59
83
1.59
99
1.61
97
1.61
100
1.60
97
1.60
99
1.62
96
1.59
97
1.60
96
1.60
95
1.60
94
1.59
95
1.59
96
1.60
97
1.59
97
1.59
97
1.59
90
1.59
97
MFMNet_retwo views1.81
88
1.91
81
1.71
95
1.95
90
1.70
100
1.95
99
1.70
101
1.96
99
1.74
100
1.97
100
1.72
99
1.95
99
1.71
96
1.97
96
1.71
98
1.86
97
1.62
98
1.85
98
1.66
98
1.86
92
1.64
98
FBW_ROBtwo views2.12
89
2.46
85
1.77
96
2.49
97
1.79
101
2.38
103
1.83
102
2.46
103
1.78
101
2.48
104
1.97
100
2.40
103
1.78
97
2.42
100
1.83
100
2.31
101
1.85
99
2.38
103
1.82
99
2.35
97
1.84
99
ccstwo views2.14
90
7.84
93
0.42
60
8.04
106
0.27
58
0.44
65
0.30
60
0.33
57
0.42
65
0.41
64
0.42
64
0.63
68
6.74
109
0.45
62
0.22
50
0.53
65
0.37
60
0.36
62
0.27
56
7.74
107
6.56
108
NCCL2two views2.28
91
2.27
84
2.28
97
2.28
95
2.28
102
2.27
101
2.29
103
2.28
101
2.28
102
2.27
102
2.28
102
2.28
101
2.27
98
2.27
98
2.28
103
2.28
100
2.28
101
2.27
101
2.29
101
2.27
95
2.29
100
Abc-Nettwo views2.32
92
24.75
100
1.14
83
1.17
77
1.17
93
1.11
90
1.15
91
1.17
90
1.21
94
1.08
85
1.16
90
1.12
91
1.14
85
1.09
83
1.17
89
1.09
84
1.16
93
1.17
92
1.17
91
1.16
84
1.11
86
stereogantwo views2.33
93
24.38
99
1.18
88
1.18
78
1.18
95
1.14
92
1.18
95
1.19
93
1.19
93
1.14
93
1.18
94
1.14
92
1.18
89
1.14
89
1.18
90
1.14
93
1.14
91
1.19
95
1.19
92
1.15
83
1.19
93
RPtwo views2.33
93
25.00
101
1.13
82
1.15
74
1.15
89
1.10
87
1.15
91
1.15
87
1.15
88
1.12
91
1.15
87
1.09
85
1.13
83
1.10
85
1.15
86
1.11
91
1.09
86
1.16
88
1.12
87
1.16
84
1.17
91
Anonymous Stereotwo views2.36
95
18.47
97
11.50
109
9.80
109
0.53
68
0.28
53
0.50
67
0.58
67
0.53
67
0.59
68
0.53
67
0.66
69
0.52
63
0.62
65
0.41
61
0.36
60
0.37
60
0.31
57
0.19
52
0.20
49
0.22
53
NCC-stereotwo views2.36
95
25.52
103
1.17
86
1.15
74
1.17
93
1.11
90
1.11
89
1.15
87
1.16
89
1.10
86
1.16
90
1.10
87
1.15
86
1.14
89
1.13
84
1.09
84
1.11
88
1.15
87
1.16
90
1.17
89
1.17
91
RGCtwo views2.36
95
25.48
102
1.19
89
1.15
74
1.15
89
1.10
87
1.16
93
1.17
90
1.16
89
1.12
91
1.16
90
1.11
90
1.13
83
1.09
83
1.18
90
1.10
86
1.15
92
1.16
88
1.12
87
1.14
82
1.12
88
AF-Nettwo views2.37
98
25.71
104
1.17
86
1.13
72
1.15
89
1.15
93
1.18
95
1.19
93
1.16
89
1.10
86
1.15
87
1.10
87
1.18
89
1.12
88
1.15
86
1.10
86
1.10
87
1.17
92
1.15
89
1.16
84
1.12
88
Nwc_Nettwo views2.37
98
25.95
105
1.15
85
1.14
73
1.15
89
1.08
85
1.14
90
1.17
90
1.16
89
1.14
93
1.15
87
1.08
84
1.15
86
1.11
87
1.14
85
1.10
86
1.11
88
1.16
88
1.09
86
1.16
84
1.15
90
CFNet_RVCtwo views2.38
100
8.71
95
0.38
58
9.33
108
0.42
65
0.56
68
0.42
65
0.60
68
0.38
63
0.62
69
0.42
64
0.56
66
7.29
110
0.62
65
0.42
62
0.53
65
0.37
60
0.49
66
0.34
61
8.30
108
6.87
109
edge stereotwo views2.43
101
27.07
106
1.14
83
1.06
69
1.14
88
1.08
85
1.17
94
1.16
89
1.14
87
1.10
86
1.16
90
1.09
85
1.16
88
1.10
85
1.16
88
1.10
86
1.11
88
1.16
88
1.19
92
1.13
81
1.11
86
DPSNettwo views3.66
102
3.60
88
3.62
104
3.63
100
3.64
107
3.65
106
3.65
107
3.66
106
3.67
106
3.67
106
3.65
107
3.67
106
3.66
102
3.66
104
3.68
107
3.68
105
3.66
105
3.67
106
3.68
105
3.67
101
3.67
105
STTStereotwo views3.73
103
30.40
107
2.37
98
2.39
96
2.31
103
2.35
102
2.33
104
2.29
102
2.42
103
2.32
103
2.34
103
2.33
102
2.36
99
2.30
99
2.27
102
2.35
102
2.22
100
2.31
102
2.22
100
2.29
96
2.34
101
DispFullNettwo views4.61
104
4.84
89
3.12
102
5.00
101
3.57
106
4.75
107
3.03
106
7.75
108
4.17
107
4.91
107
3.09
106
6.20
107
3.95
103
6.71
106
3.51
106
5.30
106
3.67
106
5.86
108
3.63
104
5.61
105
3.46
104
CC-Net-ROBtwo views4.74
105
40.54
109
2.83
101
2.84
98
2.86
105
2.81
105
2.90
105
2.91
105
2.90
105
2.84
105
2.89
105
2.78
104
2.83
101
2.82
102
2.90
105
2.80
103
2.80
103
2.89
105
2.85
103
2.88
100
2.88
103
NLCA_NET_v2_RVCtwo views5.04
106
50.87
110
2.78
100
2.87
99
2.71
104
2.78
104
0.47
66
2.88
104
2.86
104
1.89
98
2.67
104
2.79
105
2.79
100
2.78
101
2.83
104
2.82
104
2.82
104
2.83
104
2.77
102
2.83
99
2.76
102
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
SGM-Foresttwo views5.21
107
5.92
92
4.08
105
6.18
105
4.16
108
6.31
108
4.34
108
6.50
107
4.33
108
6.14
108
4.21
108
6.61
108
4.55
106
6.67
105
4.48
108
5.94
108
3.94
107
5.85
107
4.03
106
5.79
106
4.17
106
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
108
9.15
96
5.38
108
8.84
107
5.18
109
8.80
109
5.31
109
8.79
109
5.28
109
8.89
109
5.20
109
8.93
109
5.33
108
8.95
109
5.36
109
8.70
109
5.21
109
8.74
109
5.20
109
8.89
109
5.15
107
PA-Nettwo views11.80
109
223.51
114
0.62
63
0.59
51
0.71
74
0.59
69
0.73
74
0.67
69
0.73
73
0.55
67
0.61
69
0.60
67
0.74
69
0.63
67
0.73
69
0.66
68
0.60
67
0.69
71
0.66
68
0.72
65
0.65
66
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
MANEtwo views19.05
110
23.00
98
15.00
110
23.00
110
15.00
110
24.00
110
16.00
110
24.00
110
16.00
110
22.00
110
15.00
110
23.00
110
16.00
111
23.00
110
15.00
110
22.00
110
15.00
110
22.00
110
15.00
110
22.00
110
15.00
110
WCMA_ROBtwo views31.10
111
35.43
108
27.12
111
39.51
111
23.10
111
38.78
111
25.30
111
37.49
111
25.39
111
37.29
111
27.02
111
38.52
111
26.48
112
37.80
111
26.44
111
36.28
111
22.65
111
33.90
111
22.30
111
37.10
111
24.18
111
MDST_ROBtwo views69.83
112
87.70
111
41.95
112
113.75
114
65.62
114
75.05
112
55.25
113
75.64
113
45.04
112
71.61
112
41.75
112
72.81
112
44.06
113
68.38
112
44.63
112
101.89
114
59.57
114
107.10
114
61.05
114
104.38
114
59.38
114
NOSS_ROBtwo views102.95
113
153.00
112
121.00
114
51.00
112
44.00
112
165.00
115
127.00
114
153.00
114
119.00
114
164.00
115
125.00
114
168.00
115
120.00
115
153.00
114
117.00
114
49.00
112
44.00
112
49.00
112
44.00
112
49.00
112
44.00
112
CBMVpermissivetwo views128.50
114
1422.70
119
53.10
113
79.50
113
51.30
113
77.30
113
49.70
112
74.00
112
48.20
113
77.20
113
48.80
113
73.90
113
48.00
114
73.80
113
48.80
113
70.40
113
45.10
113
68.90
113
46.10
113
68.20
113
45.00
113
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
115
171.00
113
160.68
115
162.58
115
160.59
115
164.01
114
160.35
115
158.51
115
158.56
115
158.34
114
160.12
115
158.56
114
159.92
116
157.26
115
158.94
115
154.38
115
158.36
115
155.75
115
159.13
115
153.67
115
154.07
115
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
116
354.61
115
207.27
116
363.24
116
206.46
116
364.72
116
210.41
116
364.72
116
210.41
116
364.81
116
208.64
116
364.81
116
208.64
117
364.72
116
210.41
116
354.70
116
205.53
116
354.70
116
205.53
116
354.70
116
205.53
116
LE_ROBtwo views396.57
117
471.28
116
329.84
117
471.48
117
308.15
117
526.83
117
322.10
117
488.15
117
323.76
117
495.46
117
317.97
117
497.17
117
320.10
118
481.62
117
326.76
117
462.71
117
298.97
117
466.16
117
285.98
117
447.62
117
289.21
117
SGM-ForestMtwo views596.69
118
677.77
117
444.52
118
699.85
118
517.25
118
732.94
118
488.29
118
770.79
118
460.11
118
750.81
118
487.98
118
792.79
118
499.41
119
730.90
118
475.81
118
720.03
118
491.16
118
663.96
118
418.60
118
674.76
118
436.05
118
CBMV_ROBtwo views818.48
119
913.88
118
709.52
119
862.84
119
597.78
119
1073.99
119
700.52
119
1015.66
119
702.59
119
1115.65
119
760.02
119
1130.24
119
721.57
120
1037.41
119
692.65
119
814.05
119
564.29
119
843.28
119
595.31
119
915.51
119
602.92
119
DPM-Stereotwo views10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
121
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
10000000.00
120
MSMDNettwo views5.27
107