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
2
0.63
7
0.54
10
0.36
13
0.69
3
0.54
17
1.22
3
0.72
2
1.61
55
1.50
50
0.75
4
0.53
2
0.64
1
0.44
1
0.18
2
0.15
8
0.17
3
0.15
6
0.17
2
0.17
4
R-Stereo Traintwo views0.57
1
0.26
2
0.63
7
0.54
10
0.36
13
0.69
3
0.54
17
1.22
3
0.72
2
1.61
55
1.50
50
0.75
4
0.53
2
0.64
1
0.44
1
0.18
2
0.15
8
0.17
3
0.15
6
0.17
2
0.17
4
PMTNettwo views0.57
1
0.23
1
0.44
1
0.39
1
0.27
4
1.01
27
0.58
28
1.18
1
1.35
46
0.74
1
0.83
1
0.60
1
0.53
2
1.84
32
0.55
4
0.20
10
0.14
3
0.16
2
0.11
1
0.16
1
0.14
1
DPM-Stereotwo views0.62
4
0.30
6
0.83
18
0.48
2
0.36
13
0.95
22
0.33
1
2.37
95
0.67
1
1.25
23
1.00
4
0.68
2
0.51
1
0.91
4
0.61
7
0.18
2
0.15
8
0.21
14
0.17
11
0.19
6
0.19
8
DN-CSS_ROBtwo views0.64
5
0.32
9
1.05
31
0.55
13
0.40
23
0.54
1
0.45
4
1.21
2
0.93
4
1.03
11
1.47
47
0.69
3
0.63
6
1.76
28
0.49
3
0.18
2
0.14
3
0.25
19
0.22
20
0.22
11
0.20
11
HITNettwo views0.65
6
0.40
23
0.77
15
0.48
2
0.24
2
1.05
31
0.49
10
1.76
40
1.02
8
1.18
20
1.18
15
0.89
6
0.80
16
0.85
3
0.79
20
0.17
1
0.12
1
0.21
14
0.17
11
0.29
20
0.15
2
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
ccstwo views0.69
7
0.29
4
0.64
10
0.55
13
0.32
8
0.63
2
0.45
4
2.30
92
1.43
56
0.85
2
1.02
6
1.05
12
0.75
10
0.97
6
0.69
10
0.30
38
0.25
41
0.30
32
0.29
41
0.31
28
0.35
49
AdaStereotwo views0.70
8
0.43
31
0.75
13
0.73
67
0.37
17
0.92
17
0.53
15
1.85
54
1.18
16
1.12
14
1.23
19
1.07
13
0.65
8
0.96
5
0.63
8
0.31
41
0.15
8
0.32
38
0.21
18
0.29
20
0.22
16
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.
MLCVtwo views0.72
9
0.34
13
1.10
36
0.51
4
0.23
1
0.91
14
0.39
3
1.33
5
1.23
25
1.46
44
1.55
54
1.01
10
0.64
7
1.95
39
0.80
23
0.18
2
0.12
1
0.18
7
0.14
4
0.19
6
0.17
4
CFNet_RVCtwo views0.72
9
0.32
9
0.54
3
0.66
41
0.54
45
0.83
8
0.55
22
1.37
6
1.15
13
0.93
5
1.28
24
1.57
47
0.97
32
1.35
10
0.81
25
0.26
29
0.20
24
0.33
43
0.25
24
0.30
25
0.23
18
BEATNet_4xtwo views0.73
11
0.49
46
1.17
43
0.53
7
0.26
3
1.03
28
0.55
22
1.82
51
1.06
9
1.21
21
1.22
18
0.96
9
0.91
26
1.01
7
1.06
36
0.21
13
0.14
3
0.24
18
0.19
14
0.37
42
0.21
12
DeepPruner_ROBtwo views0.74
12
0.48
44
0.87
22
0.57
17
0.44
29
0.79
7
0.54
17
1.73
35
0.96
7
1.38
38
1.27
22
1.24
22
0.73
9
1.43
11
0.73
13
0.34
53
0.28
51
0.27
21
0.26
27
0.31
28
0.31
40
DMCAtwo views0.74
12
0.39
22
0.87
22
0.65
39
0.56
49
0.74
5
0.56
25
1.37
6
0.94
5
1.42
42
1.13
8
1.35
30
0.98
35
1.31
9
0.80
23
0.31
41
0.25
41
0.38
60
0.32
52
0.32
33
0.25
22
iResNettwo views0.75
14
0.40
23
1.36
62
0.67
44
0.29
5
0.92
17
0.59
29
1.52
18
1.25
29
1.28
26
1.48
48
0.94
8
0.76
13
1.62
22
0.78
18
0.18
2
0.14
3
0.18
7
0.15
6
0.22
11
0.23
18
iResNet_ROBtwo views0.77
15
0.30
6
0.99
28
0.57
17
0.36
13
0.83
8
0.36
2
2.33
93
1.58
70
1.27
24
1.30
27
1.25
24
0.84
17
1.65
23
0.79
20
0.20
10
0.14
3
0.17
3
0.12
2
0.18
4
0.27
25
STTStereotwo views0.78
16
0.50
50
1.22
45
0.64
35
0.54
45
0.89
10
0.54
17
1.44
11
1.11
11
1.38
38
1.09
7
1.27
27
0.85
19
1.43
11
0.83
26
0.29
36
0.27
47
0.33
43
0.39
70
0.32
33
0.30
36
ccs_robtwo views0.79
17
0.32
9
0.83
18
0.60
21
0.34
10
1.03
28
0.46
6
1.81
50
1.31
37
1.12
14
1.35
34
1.18
18
0.77
14
2.63
62
0.60
6
0.23
17
0.18
20
0.27
21
0.27
30
0.23
14
0.21
12
NLCA_NET_v2_RVCtwo views0.79
17
0.43
31
1.05
31
0.64
35
0.58
57
0.95
22
0.53
15
1.53
21
1.18
16
1.37
35
1.15
12
1.43
36
0.88
22
1.61
21
0.84
27
0.25
25
0.22
31
0.27
21
0.27
30
0.29
20
0.29
31
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
17
0.43
31
1.05
31
0.63
29
0.56
49
0.94
20
0.54
17
1.51
15
1.18
16
1.37
35
1.17
13
1.42
35
0.89
23
1.69
25
0.87
29
0.25
25
0.22
31
0.27
21
0.26
27
0.29
20
0.30
36
DMCA-RVCcopylefttwo views0.79
17
0.41
26
1.25
50
0.63
29
0.50
38
0.89
10
0.61
32
1.39
9
0.95
6
1.35
34
1.19
16
1.62
50
0.84
17
1.48
13
0.86
28
0.35
58
0.27
47
0.37
57
0.32
52
0.31
28
0.25
22
StereoDRNet-Refinedtwo views0.80
21
0.41
26
0.88
24
0.62
26
0.42
25
1.33
50
0.48
9
1.52
18
1.17
15
1.45
43
1.44
45
1.44
39
0.90
25
1.26
8
1.24
43
0.20
10
0.16
14
0.27
21
0.24
21
0.25
16
0.28
29
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
iResNetv2_ROBtwo views0.80
21
0.37
18
1.71
75
0.56
15
0.37
17
0.94
20
0.64
34
1.73
35
1.34
42
1.23
22
1.42
42
1.26
25
0.92
28
1.85
34
0.56
5
0.23
17
0.15
8
0.19
9
0.14
4
0.24
15
0.19
8
FADNet_RVCtwo views0.83
23
0.51
53
1.83
77
0.54
10
0.42
25
1.19
41
0.46
6
1.51
15
1.09
10
0.89
3
1.24
20
0.91
7
0.87
21
2.11
42
1.11
38
0.27
32
0.22
31
0.31
35
0.28
36
0.43
62
0.38
58
FADNet-RVC-Resampletwo views0.84
24
0.47
40
2.10
80
0.61
24
0.45
30
1.07
33
0.61
32
1.59
25
1.20
22
1.08
12
0.98
3
1.03
11
0.98
35
2.17
44
0.72
12
0.23
17
0.22
31
0.30
32
0.32
52
0.35
38
0.34
45
DLCB_ROBtwo views0.86
25
0.40
23
0.78
16
0.68
49
0.54
45
1.11
34
0.84
47
1.49
13
1.23
25
1.64
58
1.65
57
1.70
56
0.91
26
1.70
26
1.05
35
0.25
25
0.24
39
0.26
20
0.27
30
0.27
19
0.24
20
HSM-Net_RVCpermissivetwo views0.87
26
0.29
4
0.53
2
0.53
7
0.30
7
1.15
39
0.55
22
3.57
109
1.28
36
1.53
49
1.74
64
1.68
54
0.93
30
1.49
14
0.77
17
0.22
15
0.17
16
0.19
9
0.18
13
0.20
8
0.19
8
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
FADNet-RVCtwo views0.88
27
0.67
80
1.93
79
0.60
21
0.48
34
0.91
14
0.60
31
1.71
33
1.18
16
0.90
4
0.96
2
1.36
31
0.92
28
2.25
50
1.17
40
0.30
38
0.29
55
0.32
38
0.32
52
0.41
55
0.35
49
PSMNet_ROBtwo views0.90
28
0.52
54
1.17
43
0.72
61
0.61
60
1.27
46
1.12
73
1.76
40
1.12
12
1.09
13
1.28
24
1.45
40
0.89
23
1.65
23
1.42
60
0.33
47
0.25
41
0.37
57
0.39
70
0.35
38
0.29
31
CBMV_ROBtwo views0.90
28
0.44
35
0.63
7
0.51
4
0.37
17
1.25
45
0.49
10
1.77
43
1.27
33
1.41
41
2.00
76
1.33
29
1.08
46
2.17
44
0.76
16
0.43
77
0.42
87
0.47
85
0.44
82
0.41
55
0.36
52
NVstereo2Dtwo views0.92
30
0.41
26
1.48
67
0.72
61
0.72
69
1.20
43
0.87
51
1.75
38
1.38
51
0.99
7
1.13
8
1.26
25
0.98
35
2.21
48
0.89
30
0.43
77
0.20
24
0.38
60
0.24
21
0.57
87
0.63
98
FADNettwo views0.92
30
0.61
73
1.81
76
0.62
26
0.55
48
0.96
24
0.57
26
1.78
47
1.33
41
0.93
5
1.01
5
1.36
31
1.01
40
2.71
65
0.75
14
0.33
47
0.34
68
0.31
35
0.34
60
0.59
90
0.40
59
CBMVpermissivetwo views0.95
32
0.44
35
0.66
12
0.53
7
0.35
11
1.53
60
1.63
90
1.69
30
1.42
55
2.00
77
1.60
56
1.43
36
1.17
51
1.59
19
0.99
34
0.34
53
0.34
68
0.39
66
0.33
57
0.33
37
0.32
41
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
ETE_ROBtwo views0.96
33
0.64
78
1.16
41
0.69
50
0.48
34
1.14
37
1.13
74
1.77
43
1.19
20
1.91
74
1.45
46
1.73
57
1.05
43
1.60
20
1.14
39
0.32
45
0.26
45
0.38
60
0.34
60
0.39
51
0.41
62
CFNettwo views0.96
33
0.36
16
1.01
29
0.62
26
0.42
25
0.99
26
0.47
8
1.90
62
1.32
39
0.99
7
1.39
38
1.10
15
0.86
20
5.58
89
0.79
20
0.23
17
0.17
16
0.28
27
0.27
30
0.26
18
0.21
12
XPNet_ROBtwo views0.96
33
0.49
46
1.09
35
0.72
61
0.57
51
1.11
34
1.14
75
1.66
28
1.23
25
1.85
69
1.39
38
1.64
52
1.11
48
1.80
30
1.42
60
0.38
66
0.30
57
0.33
43
0.29
41
0.37
42
0.34
45
StereoDRNettwo views0.98
36
0.47
40
1.70
74
0.76
72
0.94
93
1.51
58
0.91
56
1.92
65
1.32
39
1.61
55
1.13
8
1.47
43
0.97
32
1.75
27
1.46
63
0.28
34
0.23
36
0.29
29
0.27
30
0.35
38
0.27
25
NCCL2two views1.00
37
0.56
61
1.06
34
0.75
71
0.57
51
1.31
49
1.72
91
1.77
43
1.36
48
1.50
47
1.17
13
1.91
64
1.17
51
1.54
16
1.29
50
0.35
58
0.31
61
0.45
79
0.44
82
0.40
54
0.40
59
DRN-Testtwo views1.00
37
0.44
35
1.13
39
0.70
54
0.81
84
1.62
64
0.79
46
2.14
80
1.44
58
1.78
65
1.33
31
1.43
36
1.05
43
2.18
47
1.53
67
0.27
32
0.21
27
0.31
35
0.30
46
0.32
33
0.28
29
LALA_ROBtwo views1.01
39
0.59
66
1.16
41
0.67
44
0.51
43
1.63
65
1.26
81
1.69
30
1.38
51
1.81
67
1.37
35
1.94
69
1.02
41
1.58
18
1.38
57
0.37
63
0.27
47
0.43
76
0.37
68
0.43
62
0.37
53
PA-Nettwo views1.01
39
0.57
63
1.29
54
0.64
35
0.74
71
1.19
41
0.84
47
1.69
30
1.43
56
1.16
18
1.19
16
1.32
28
1.19
54
3.42
73
1.25
44
0.29
36
0.45
94
0.33
43
0.49
92
0.32
33
0.47
78
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
DISCOtwo views1.03
41
0.34
13
1.11
37
0.69
50
0.45
30
2.07
80
0.84
47
2.13
79
1.50
63
1.15
17
1.33
31
2.40
86
1.07
45
2.60
60
1.51
66
0.23
17
0.18
20
0.22
16
0.20
17
0.31
28
0.27
25
TDLMtwo views1.06
42
0.44
35
0.79
17
0.63
29
0.50
38
0.89
10
1.44
86
1.48
12
1.26
30
1.12
14
1.40
40
1.08
14
0.98
35
6.65
95
0.78
18
0.33
47
0.21
27
0.33
43
0.26
27
0.37
42
0.24
20
CVANet_RVCtwo views1.07
43
0.43
31
0.76
14
0.60
21
0.50
38
0.90
13
0.97
64
1.52
18
1.26
30
1.33
32
1.38
36
1.24
22
1.18
53
6.75
97
0.75
14
0.34
53
0.23
36
0.34
49
0.29
41
0.42
58
0.27
25
GANetREF_RVCpermissivetwo views1.07
43
0.86
95
1.31
56
0.82
75
0.53
44
1.41
54
1.31
83
1.76
40
1.52
65
1.27
24
1.31
29
1.46
42
1.24
56
2.03
40
0.96
33
0.56
99
0.45
94
0.67
104
0.50
93
0.88
104
0.59
93
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
45
0.61
73
1.32
58
0.85
80
0.72
69
1.37
52
0.51
13
1.54
22
1.15
13
1.78
65
1.54
53
2.01
74
1.11
48
2.90
68
2.00
81
0.33
47
0.30
57
0.40
71
0.30
46
0.44
67
0.45
72
Nwc_Nettwo views1.09
46
0.49
46
1.33
59
0.89
87
0.80
83
1.49
57
0.66
37
2.19
86
1.27
33
1.32
29
2.26
82
1.96
71
1.54
70
1.92
35
1.34
54
0.40
69
0.30
57
0.43
76
0.45
84
0.38
48
0.37
53
RPtwo views1.09
46
0.50
50
1.12
38
0.90
88
1.04
98
1.11
34
0.84
47
1.80
48
1.62
71
1.55
52
2.00
76
1.90
63
1.78
84
1.84
32
1.25
44
0.52
94
0.36
76
0.48
86
0.36
67
0.45
69
0.43
69
NCC-stereotwo views1.10
48
0.54
57
1.24
47
0.84
77
0.94
93
1.29
47
0.89
53
2.01
74
1.34
42
1.32
29
1.74
64
1.88
59
1.74
82
2.35
53
1.25
44
0.45
82
0.37
78
0.45
79
0.58
99
0.42
58
0.42
66
Abc-Nettwo views1.10
48
0.54
57
1.24
47
0.84
77
0.94
93
1.29
47
0.89
53
2.01
74
1.34
42
1.32
29
1.74
64
1.88
59
1.74
82
2.35
53
1.25
44
0.45
82
0.37
78
0.45
79
0.58
99
0.42
58
0.42
66
edge stereotwo views1.11
50
0.50
50
1.30
55
0.71
59
0.65
64
1.14
37
0.75
42
1.93
68
1.37
50
1.71
61
1.83
69
2.54
90
1.44
67
2.43
56
1.23
42
0.42
75
0.41
86
0.54
95
0.35
63
0.42
58
0.46
75
RASNettwo views1.12
51
0.30
6
0.91
26
0.66
41
0.77
79
0.97
25
0.57
26
1.37
6
1.35
46
1.02
10
1.34
33
2.11
78
1.43
66
7.09
98
1.39
59
0.19
9
0.17
16
0.17
3
0.19
14
0.18
4
0.16
3
AF-Nettwo views1.13
52
0.52
54
1.33
59
0.86
81
0.91
89
1.38
53
0.71
40
2.16
83
1.62
71
1.56
53
2.40
87
1.99
73
1.65
78
1.81
31
1.48
64
0.48
88
0.28
51
0.49
88
0.33
57
0.39
51
0.32
41
NaN_ROBtwo views1.13
52
0.55
59
1.15
40
0.64
35
0.48
34
1.61
63
2.03
102
1.72
34
1.48
60
2.09
78
1.27
22
1.40
34
1.08
46
3.94
77
1.30
51
0.28
34
0.34
68
0.28
27
0.28
36
0.30
25
0.33
44
RGCtwo views1.14
54
0.79
91
1.26
51
0.91
90
0.96
96
1.52
59
0.77
43
1.92
65
1.26
30
1.54
50
1.93
74
2.02
75
1.54
70
2.25
50
1.31
52
0.44
81
0.35
73
0.58
102
0.48
90
0.47
71
0.46
75
stereogantwo views1.15
55
0.42
29
1.40
63
0.86
81
0.84
86
2.38
87
0.71
40
1.88
58
1.73
79
1.72
62
1.94
75
2.31
81
1.29
61
2.07
41
1.08
37
0.36
60
0.35
73
0.39
66
0.31
49
0.51
81
0.46
75
AANet_RVCtwo views1.16
56
0.52
54
1.49
68
0.65
39
0.45
30
0.75
6
0.69
39
1.42
10
1.55
69
1.34
33
1.87
72
1.53
45
0.78
15
7.43
100
1.38
57
0.25
25
0.17
16
0.19
9
0.16
9
0.22
11
0.26
24
RYNettwo views1.17
57
0.49
46
1.24
47
0.70
54
0.97
97
1.71
67
0.92
59
1.91
63
1.46
59
1.38
38
1.14
11
1.88
59
1.30
62
4.24
80
1.87
74
0.26
29
0.21
27
0.35
51
0.27
30
0.55
85
0.54
87
HSMtwo views1.19
58
0.32
9
0.65
11
0.57
17
0.38
21
1.23
44
0.59
29
2.25
89
1.19
20
1.30
28
1.25
21
1.94
69
1.03
42
9.32
107
0.71
11
0.22
15
0.16
14
0.20
12
0.16
9
0.21
10
0.21
12
SGM-Foresttwo views1.24
59
0.38
20
0.60
4
0.52
6
0.37
17
1.56
62
1.30
82
1.65
27
1.36
48
2.11
79
1.79
67
1.63
51
0.93
30
7.48
101
0.95
32
0.36
60
0.37
78
0.39
66
0.34
60
0.37
42
0.32
41
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
59
0.76
90
2.46
88
0.72
61
0.78
82
1.15
39
0.49
10
1.75
38
1.92
89
2.31
82
3.07
94
1.59
48
1.40
64
2.46
57
2.10
85
0.33
47
0.22
31
0.35
51
0.24
21
0.38
48
0.41
62
NOSS_ROBtwo views1.26
61
0.46
39
0.62
6
0.56
15
0.38
21
1.06
32
0.78
44
1.73
35
1.40
53
1.01
9
1.43
44
1.10
15
0.75
10
10.62
112
0.65
9
0.45
82
0.43
88
0.48
86
0.45
84
0.44
67
0.41
62
Anonymous Stereotwo views1.28
62
0.75
88
2.45
87
0.74
69
0.77
79
0.93
19
1.77
93
1.51
15
1.31
37
1.67
59
1.29
26
1.20
20
0.75
10
6.29
92
1.68
71
0.37
63
0.36
76
0.38
60
0.40
74
0.43
62
0.45
72
RTSCtwo views1.29
63
0.73
83
2.44
86
0.83
76
0.57
51
1.85
74
0.65
35
1.88
58
2.00
91
2.20
80
1.66
58
1.60
49
1.26
58
2.33
52
3.92
95
0.32
45
0.23
36
0.27
21
0.29
41
0.37
42
0.37
53
PWCDC_ROBbinarytwo views1.29
63
0.75
88
1.33
59
1.05
99
0.88
88
1.73
68
0.52
14
1.89
60
3.28
103
1.29
27
4.88
102
1.22
21
1.46
69
1.77
29
1.45
62
0.60
101
0.26
45
0.29
29
0.25
24
0.52
82
0.37
53
SHDtwo views1.31
65
0.71
82
2.34
84
0.86
81
0.75
76
1.77
70
0.65
35
2.22
88
2.50
98
2.40
85
1.71
62
1.91
64
1.60
75
1.93
38
2.37
87
0.38
66
0.31
61
0.38
60
0.40
74
0.48
73
0.50
83
PDISCO_ROBtwo views1.32
66
0.65
79
1.44
64
1.36
108
1.46
111
2.55
90
0.92
59
2.45
98
2.07
93
1.37
35
1.49
49
1.65
53
1.00
39
2.82
66
1.35
56
1.01
108
0.27
47
0.54
95
0.47
88
0.71
103
0.73
104
PASMtwo views1.34
67
0.92
98
2.62
91
0.86
81
0.87
87
1.03
28
1.08
69
1.50
14
1.51
64
1.82
68
1.41
41
1.68
54
1.26
58
4.87
84
1.34
54
0.61
102
0.65
107
0.68
105
0.78
106
0.69
101
0.68
103
ADCReftwo views1.35
68
0.55
59
2.37
85
0.72
61
0.77
79
1.94
76
0.78
44
1.91
63
1.27
33
1.57
54
1.59
55
1.36
31
1.41
65
1.54
16
7.53
104
0.23
17
0.20
24
0.32
38
0.30
46
0.29
20
0.29
31
GANettwo views1.36
69
0.47
40
0.86
20
0.67
44
0.45
30
1.34
51
1.73
92
1.64
26
1.24
28
1.49
46
2.38
86
1.55
46
1.20
55
9.30
106
0.93
31
0.30
38
0.31
61
0.33
43
0.28
36
0.39
51
0.30
36
XQCtwo views1.37
70
0.87
96
2.56
89
0.84
77
0.69
68
1.83
73
0.89
53
2.07
78
1.88
86
1.88
70
1.53
52
1.91
64
1.55
72
2.85
67
3.05
90
0.48
88
0.29
55
0.44
78
0.39
70
0.69
101
0.60
96
MFMNet_retwo views1.40
71
1.14
105
1.68
73
1.39
109
1.07
99
1.48
56
1.43
85
1.87
56
1.82
84
2.26
81
2.19
81
1.51
44
1.86
89
1.52
15
1.25
44
0.93
107
0.83
109
0.84
108
0.80
107
0.99
107
1.05
109
CSANtwo views1.41
72
0.59
66
1.26
51
0.70
54
0.50
38
1.76
69
1.78
94
1.77
43
1.90
88
2.35
83
2.32
84
1.92
67
1.94
91
5.56
88
1.49
65
0.43
77
0.35
73
0.40
71
0.41
76
0.43
62
0.41
62
MFN_U_SF_RVCtwo views1.52
73
0.84
92
3.42
94
0.87
85
0.66
65
2.56
91
0.95
62
2.36
94
1.70
76
1.98
76
1.86
71
2.45
88
1.79
85
2.95
69
2.58
88
0.52
94
0.48
98
0.55
98
0.59
102
0.59
90
0.61
97
SAMSARAtwo views1.55
74
0.84
92
2.61
90
1.22
107
1.07
99
2.50
89
1.93
99
2.05
77
1.89
87
2.75
92
1.66
58
2.81
95
1.69
80
3.08
70
1.97
80
0.46
86
0.51
102
0.42
75
0.46
87
0.58
88
0.59
93
DeepPrunerFtwo views1.60
75
0.73
83
5.81
104
0.88
86
1.19
107
0.91
14
0.95
62
2.19
86
5.24
109
1.17
19
1.30
27
1.19
19
1.30
62
4.90
85
1.66
70
0.42
75
0.34
68
0.52
91
0.47
88
0.41
55
0.40
59
STTStereo_v2two views1.66
76
0.63
76
1.56
70
1.11
103
0.74
71
2.98
100
0.91
56
1.55
23
1.22
23
2.76
93
8.18
109
2.32
82
2.53
99
1.92
35
1.87
74
0.48
88
0.44
90
0.45
79
0.31
49
0.60
93
0.57
91
G-Nettwo views1.66
76
0.63
76
1.56
70
1.11
103
0.74
71
2.98
100
0.91
56
1.55
23
1.22
23
2.76
93
8.18
109
2.32
82
2.53
99
1.92
35
1.87
74
0.48
88
0.44
90
0.45
79
0.31
49
0.60
93
0.57
91
MDST_ROBtwo views1.67
78
0.34
13
1.58
72
0.92
92
0.58
57
4.26
108
1.11
71
2.83
102
1.41
54
4.49
104
2.55
89
1.45
40
0.97
32
7.79
103
1.17
40
0.31
41
0.25
41
0.37
57
0.33
57
0.31
28
0.30
36
SPS-STEREOcopylefttwo views1.67
78
1.04
100
1.23
46
1.21
106
1.13
105
1.82
72
1.11
71
2.16
83
1.52
65
3.07
100
2.32
84
2.82
96
1.80
86
4.11
79
2.04
82
1.03
110
0.99
112
0.97
109
0.94
108
1.03
109
1.02
107
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
SGM_RVCbinarytwo views1.68
80
0.36
16
0.61
5
0.59
20
0.29
5
2.25
85
1.18
78
2.38
96
1.68
75
3.60
101
2.67
90
3.69
101
2.42
97
8.13
104
1.90
77
0.31
41
0.31
61
0.30
32
0.28
36
0.30
25
0.29
31
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DPSNettwo views1.69
81
0.59
66
4.33
98
0.70
54
0.57
51
2.02
79
1.62
89
3.35
108
1.78
81
1.48
45
1.38
36
2.17
79
2.84
101
4.77
83
3.66
94
0.48
88
0.31
61
0.29
29
0.25
24
0.64
97
0.48
80
PVDtwo views1.70
82
0.84
92
2.26
83
0.94
95
0.83
85
2.48
88
0.97
64
2.59
100
3.35
104
3.01
98
2.77
91
2.34
84
2.47
98
2.67
64
3.14
92
0.47
87
0.45
94
0.59
103
0.48
90
0.59
90
0.75
105
FBW_ROBtwo views1.71
83
0.62
75
1.45
65
0.67
44
0.61
60
1.81
71
1.08
69
2.52
99
1.70
76
1.77
64
1.73
63
1.98
72
1.25
57
12.57
114
1.84
73
0.34
53
0.37
78
0.56
100
0.38
69
0.45
69
0.43
69
ADCLtwo views1.72
84
0.57
63
3.66
96
0.67
44
0.61
60
2.61
93
1.85
97
1.96
70
1.67
74
1.90
73
1.83
69
2.06
77
1.90
90
2.13
43
9.15
105
0.26
29
0.24
39
0.34
49
0.32
52
0.38
48
0.35
49
MFN_U_SF_DS_RVCtwo views1.74
85
0.87
96
3.56
95
0.80
73
0.75
76
3.72
104
2.44
107
1.82
51
1.82
84
1.74
63
1.91
73
1.92
67
1.26
58
7.69
102
1.28
49
0.45
82
0.47
97
0.52
91
0.77
105
0.50
77
0.50
83
WCMA_ROBtwo views1.84
86
0.47
40
1.31
56
0.69
50
0.57
51
2.11
81
1.14
75
1.87
56
1.73
79
4.04
102
6.43
105
5.62
106
3.61
104
2.59
59
1.94
78
0.52
94
0.44
90
0.39
66
0.42
78
0.48
73
0.47
78
ADCPNettwo views1.85
87
0.69
81
8.07
106
0.71
59
0.58
57
2.79
96
1.22
80
1.94
69
1.54
68
1.88
70
1.69
60
2.46
89
2.04
93
2.37
55
6.48
102
0.38
66
0.37
78
0.38
60
0.43
80
0.47
71
0.45
72
NVStereoNet_ROBtwo views1.88
88
1.07
101
1.45
65
1.07
102
1.17
106
1.47
55
1.48
87
1.89
60
1.93
90
1.94
75
4.67
100
3.08
98
1.80
86
6.51
94
1.65
69
1.06
112
0.86
110
1.08
111
1.32
111
0.93
105
1.26
111
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
SANettwo views1.88
88
0.57
63
2.18
82
0.63
29
0.40
23
2.80
97
1.57
88
2.39
97
5.88
110
2.36
84
3.57
96
3.72
102
2.97
102
4.74
82
1.76
72
0.34
53
0.30
57
0.32
38
0.29
41
0.48
73
0.37
53
ADCMidtwo views1.89
90
0.73
83
5.06
100
0.72
61
0.67
66
2.00
78
1.14
75
1.97
72
1.48
60
2.86
96
1.81
68
2.37
85
2.05
94
2.24
49
10.39
107
0.33
47
0.31
61
0.39
66
0.43
80
0.50
77
0.44
71
AnyNet_C32two views1.96
91
1.10
103
5.17
101
0.90
88
1.11
102
2.12
82
2.37
105
1.85
54
1.48
60
2.45
86
1.70
61
2.05
76
1.64
77
2.50
58
10.28
106
0.37
63
0.34
68
0.35
51
0.41
76
0.49
76
0.54
87
MSMD_ROBtwo views1.99
92
0.59
66
0.89
25
0.70
54
0.61
60
3.95
107
0.94
61
1.80
48
1.79
82
4.41
103
4.43
98
8.17
108
3.66
105
3.40
71
1.33
53
0.52
94
0.48
98
0.54
95
0.58
99
0.53
83
0.50
83
ADCP+two views1.99
92
0.48
44
7.73
105
0.66
41
0.91
89
2.30
86
1.05
68
1.82
51
1.34
42
1.54
50
1.32
30
1.74
58
1.61
76
2.17
44
13.37
110
0.24
23
0.21
27
0.32
38
0.28
36
0.37
42
0.29
31
MSC_U_SF_DS_RVCtwo views2.12
94
1.15
106
3.10
93
1.05
99
0.76
78
5.03
110
2.40
106
2.16
83
2.62
100
1.89
72
2.28
83
2.69
93
1.65
78
9.35
108
2.04
82
0.64
105
0.51
102
0.68
105
0.71
104
1.09
110
0.67
102
ADCStwo views2.25
95
0.99
99
5.55
102
0.81
74
0.74
71
2.19
84
1.80
95
2.27
90
2.14
94
2.80
95
2.13
80
2.40
86
2.03
92
2.64
63
13.69
111
0.41
72
0.39
84
0.40
71
0.45
84
0.58
88
0.56
90
MeshStereopermissivetwo views2.26
96
0.56
61
0.92
27
0.61
24
0.49
37
2.60
92
1.02
66
3.18
106
2.05
92
7.40
112
4.98
103
7.58
107
2.98
103
6.37
93
1.96
79
0.43
77
0.43
88
0.45
79
0.39
70
0.43
62
0.42
66
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
96
1.78
109
10.22
107
0.93
93
0.74
71
2.92
99
2.20
103
2.29
91
1.80
83
2.53
88
2.41
88
2.71
94
1.73
81
3.41
72
6.72
103
0.41
72
0.39
84
0.40
71
0.42
78
0.67
100
0.63
98
DispFullNettwo views2.34
98
2.85
111
3.66
96
3.08
114
2.27
112
1.97
77
0.67
38
2.00
73
1.70
76
2.52
87
2.00
76
2.23
80
5.86
109
3.44
75
1.58
68
1.01
108
0.32
67
2.64
113
1.94
113
2.85
112
2.24
112
pmcnntwo views2.37
99
0.42
29
2.71
92
0.63
29
0.43
28
1.54
61
1.19
79
2.01
74
1.53
67
2.64
91
7.60
107
18.07
115
1.45
68
3.44
75
2.78
89
0.18
2
0.15
8
0.15
1
0.12
2
0.20
8
0.18
7
ELAScopylefttwo views2.37
99
0.74
86
1.54
69
0.99
96
0.92
91
4.78
109
2.33
104
3.33
107
3.37
105
5.05
107
4.77
101
3.67
100
4.53
107
3.42
73
4.50
97
0.62
103
0.53
104
0.55
98
0.53
96
0.63
95
0.64
101
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SGM+DAISYtwo views2.41
101
1.07
101
2.14
81
1.18
105
1.10
101
2.14
83
1.36
84
1.96
70
1.66
73
5.46
108
7.07
106
4.95
105
4.32
106
5.50
87
2.26
86
1.04
111
1.02
113
0.99
110
0.97
109
1.02
108
1.04
108
ELAS_RVCcopylefttwo views2.48
102
0.74
86
1.90
78
1.05
99
0.92
91
2.76
94
4.01
111
2.76
101
4.11
107
4.69
105
4.63
99
4.62
103
5.96
110
3.97
78
3.96
96
0.63
104
0.53
104
0.56
100
0.52
95
0.64
97
0.63
98
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SuperBtwo views2.53
103
0.59
66
18.45
112
0.69
50
0.50
38
1.87
75
0.88
52
1.67
29
2.39
96
1.50
47
1.42
42
1.14
17
1.15
50
4.28
81
5.07
98
0.24
23
0.19
22
0.22
16
0.19
14
7.86
117
0.34
45
RTStwo views2.67
104
1.38
107
15.28
110
0.99
96
1.26
109
3.78
105
1.96
100
2.15
81
2.62
100
2.61
89
2.85
92
2.60
91
1.59
73
6.27
90
5.71
100
0.40
69
0.28
51
0.35
51
0.35
63
0.50
77
0.48
80
RTSAtwo views2.67
104
1.38
107
15.28
110
0.99
96
1.26
109
3.78
105
1.96
100
2.15
81
2.62
100
2.61
89
2.85
92
2.60
91
1.59
73
6.27
90
5.71
100
0.40
69
0.28
51
0.35
51
0.35
63
0.50
77
0.48
80
MADNet+two views2.69
106
3.22
113
11.53
108
1.44
111
1.12
104
3.19
102
1.81
96
2.85
103
2.37
95
1.68
60
2.07
79
2.85
97
2.18
95
7.41
99
5.23
99
0.79
106
0.76
108
0.69
107
0.65
103
0.98
106
0.95
106
FC-DCNNcopylefttwo views2.72
107
0.59
66
0.86
20
0.73
67
0.57
51
2.91
98
1.04
67
3.01
105
2.60
99
6.83
109
7.64
108
9.92
110
6.33
111
5.21
86
3.14
92
0.53
98
0.48
98
0.51
90
0.50
93
0.53
83
0.51
86
PWCKtwo views2.88
108
2.05
110
5.72
103
2.08
112
1.11
102
3.55
103
3.90
110
2.86
104
3.54
106
3.03
99
3.63
97
3.10
99
2.31
96
8.15
105
3.09
91
2.12
113
0.91
111
1.75
112
0.98
110
2.57
111
1.18
110
MADNet++two views5.01
109
3.14
112
4.91
99
4.55
115
3.94
113
5.20
111
4.41
112
4.40
111
4.13
108
4.95
106
6.08
104
4.74
104
5.37
108
9.40
109
10.67
108
4.65
115
4.15
114
3.97
114
3.34
114
4.09
114
4.08
113
MANEtwo views6.19
110
0.59
66
1.03
30
0.93
93
1.22
108
13.38
115
2.69
108
12.81
116
17.94
117
8.21
113
10.82
113
14.72
113
11.68
112
11.39
113
11.90
109
0.59
100
0.50
101
0.52
91
1.76
112
0.55
85
0.54
87
LSMtwo views6.33
111
1.10
103
13.19
109
1.41
110
68.17
121
1.64
66
1.90
98
1.92
65
2.46
97
2.99
97
3.11
95
1.88
59
1.80
86
2.61
61
2.06
84
0.41
72
0.58
106
0.50
89
0.57
98
0.64
97
17.66
119
SGM-ForestMtwo views6.52
112
0.38
20
1.26
51
0.63
29
0.35
11
13.59
116
5.32
113
6.32
113
6.70
112
11.14
114
9.18
112
22.51
118
13.93
113
16.89
116
20.07
113
0.36
60
0.37
78
0.36
56
0.35
63
0.36
41
0.34
45
LE_ROBtwo views8.05
113
0.37
18
23.53
115
0.74
69
0.33
9
2.77
95
2.87
109
3.71
110
27.24
118
25.39
118
8.87
111
8.46
109
26.71
118
6.65
95
21.99
114
0.21
13
0.19
22
0.20
12
0.21
18
0.25
16
0.22
16
DPSimNet_ROBtwo views11.20
114
4.98
114
35.43
118
2.90
113
4.52
114
7.89
112
17.09
114
4.66
112
6.06
111
14.89
117
10.93
114
17.29
114
36.81
120
13.18
115
15.68
112
2.86
114
6.29
115
5.75
117
5.09
115
3.91
113
7.70
114
DGTPSM_ROBtwo views13.01
115
8.23
115
21.87
113
9.51
116
18.10
115
11.56
113
32.01
118
10.63
114
16.20
113
7.03
110
19.48
117
10.15
111
16.66
114
9.83
110
22.42
115
4.75
116
8.68
116
5.37
115
10.25
116
6.16
115
11.36
115
DPSMNet_ROBtwo views13.02
116
8.23
115
21.91
114
9.53
117
18.11
116
11.56
113
32.02
119
10.64
115
16.23
114
7.05
111
19.49
118
10.16
112
16.66
114
9.85
111
22.43
116
4.77
117
8.68
116
5.38
116
10.26
117
6.17
116
11.36
115
DPSMtwo views19.08
117
19.14
118
25.17
116
22.15
118
20.85
117
28.17
117
37.82
120
26.56
117
16.85
115
14.14
115
14.84
115
21.70
116
18.75
116
21.87
117
26.55
117
10.16
118
9.30
118
10.47
118
10.42
118
14.03
118
12.61
117
DPSM_ROBtwo views19.08
117
19.14
118
25.17
116
22.15
118
20.85
117
28.17
117
37.82
120
26.56
117
16.85
115
14.14
115
14.84
115
21.70
116
18.75
116
21.87
117
26.55
117
10.16
118
9.30
118
10.47
118
10.42
118
14.03
118
12.61
117
BEATNet-Init1two views26.87
119
13.43
117
70.68
122
0.91
90
0.68
67
57.02
122
21.06
115
32.82
119
32.67
119
52.12
122
53.61
122
66.05
122
49.56
123
48.79
122
34.71
119
0.49
93
0.44
90
0.53
94
0.53
96
0.63
95
0.59
93
AVERAGE_ROBtwo views40.43
120
42.37
121
42.73
119
36.96
121
35.58
120
30.50
119
27.61
116
36.71
121
38.39
121
44.91
120
44.36
120
37.45
120
36.70
119
41.88
119
40.35
120
45.43
121
46.81
121
47.11
121
46.35
121
42.51
121
43.87
121
MEDIAN_ROBtwo views42.00
121
44.40
122
44.29
120
35.15
120
34.95
119
31.28
120
29.60
117
35.88
120
40.02
122
47.20
121
46.64
121
37.37
119
38.48
122
44.35
121
42.45
121
48.03
122
49.39
122
49.71
122
49.03
122
45.14
122
46.65
122
LSM0two views42.82
122
38.75
120
50.81
121
47.07
122
109.38
122
57.00
121
76.02
122
53.74
122
33.71
120
29.02
119
30.41
119
43.87
121
37.22
121
43.91
120
53.54
122
20.52
120
18.62
120
21.10
120
21.11
120
28.41
120
42.22
120
MSMDNettwo views0.58
5