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-Stereo Traintwo views0.44
1
0.09
4
0.94
2
0.02
10
0.00
1
1.24
3
0.09
28
3.65
27
0.11
1
1.87
30
0.42
23
0.06
2
0.01
1
0.25
2
0.05
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.02
25
R-Stereotwo views0.44
1
0.09
4
0.94
2
0.02
10
0.00
1
1.24
3
0.09
28
3.65
27
0.11
1
1.87
30
0.42
23
0.06
2
0.01
1
0.25
2
0.05
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.02
25
DPM-Stereotwo views0.58
3
0.23
17
0.90
1
0.08
32
0.00
1
1.43
19
0.01
7
6.01
86
1.38
24
0.92
15
0.11
4
0.04
1
0.19
15
0.12
1
0.07
3
0.00
1
0.01
58
0.00
1
0.00
1
0.03
17
0.01
9
PMTNettwo views0.62
4
0.17
10
1.33
10
0.00
1
0.02
28
2.50
53
0.00
1
2.42
2
1.12
11
0.34
3
0.13
5
0.06
2
0.02
3
4.08
22
0.20
4
0.00
1
0.00
1
0.00
1
0.00
1
0.03
17
0.00
1
AdaStereotwo views0.65
5
0.09
4
1.12
5
0.02
10
0.00
1
1.45
21
0.23
49
4.38
54
1.44
33
1.17
22
0.37
21
1.56
20
0.18
13
0.56
4
0.37
9
0.00
1
0.00
1
0.00
1
0.00
1
0.01
6
0.00
1
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.
DN-CSS_ROBtwo views0.77
6
0.90
79
2.01
26
0.85
90
0.00
1
1.47
23
0.01
7
2.92
6
0.93
7
0.12
1
0.53
27
0.44
6
0.16
11
4.33
27
0.35
7
0.00
1
0.00
1
0.00
1
0.00
1
0.30
88
0.01
9
CFNet_RVCtwo views0.77
6
0.46
33
1.34
11
0.01
5
0.27
71
1.32
8
0.02
11
2.40
1
0.68
5
0.63
9
0.18
8
3.23
53
0.21
18
3.97
21
0.54
19
0.00
1
0.00
1
0.01
72
0.00
1
0.03
17
0.02
25
HITNettwo views0.80
8
0.58
50
2.14
29
0.69
86
0.00
1
1.35
10
0.02
11
4.50
58
1.15
12
1.76
27
0.16
6
0.60
7
0.30
22
1.13
5
1.64
39
0.00
1
0.00
1
0.00
1
0.00
1
0.06
36
0.00
1
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
StereoDRNet-Refinedtwo views0.83
9
0.12
8
1.18
7
0.02
10
0.01
21
1.31
7
0.00
1
3.33
15
1.07
8
1.92
33
1.24
44
2.48
40
0.18
13
1.86
7
1.67
40
0.00
1
0.00
1
0.00
1
0.00
1
0.14
55
0.05
44
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
STTStereotwo views0.83
9
0.34
22
2.44
34
0.24
54
0.15
58
1.41
17
0.10
33
3.02
7
1.55
41
1.34
24
0.17
7
1.77
24
0.45
31
3.42
12
0.30
5
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.01
9
DMCAtwo views0.85
11
0.51
41
1.62
15
0.02
10
0.19
62
1.14
1
0.09
28
2.48
3
0.60
4
2.95
49
0.29
13
2.05
31
0.82
44
3.29
11
0.75
26
0.01
65
0.05
80
0.04
85
0.01
66
0.08
40
0.04
40
DeepPruner_ROBtwo views0.86
12
0.50
40
1.80
20
0.01
5
0.00
1
1.57
30
0.17
40
3.48
23
0.42
3
3.31
56
0.02
1
2.09
32
0.17
12
3.01
10
0.61
22
0.00
1
0.00
1
0.00
1
0.00
1
0.12
53
0.01
9
ccstwo views0.89
13
0.16
9
1.12
5
0.02
10
0.02
28
1.53
27
0.24
51
6.55
91
2.18
58
0.94
16
0.22
10
1.52
19
0.36
25
2.04
8
0.51
17
0.03
80
0.02
68
0.12
95
0.05
80
0.11
52
0.06
51
BEATNet_4xtwo views0.97
14
0.77
69
3.36
52
0.46
74
0.01
21
1.32
8
0.09
28
4.57
60
1.26
15
1.88
32
0.22
10
1.09
16
0.42
29
1.79
6
2.09
52
0.00
1
0.00
1
0.00
1
0.00
1
0.08
40
0.06
51
iResNettwo views1.00
15
0.45
32
3.08
47
0.86
93
0.01
21
1.97
43
0.01
7
3.67
29
1.49
39
1.77
28
1.09
36
0.35
5
0.13
10
3.83
19
1.27
34
0.00
1
0.00
1
0.00
1
0.00
1
0.07
38
0.01
9
MLCVtwo views1.01
16
0.53
43
2.17
30
0.03
19
0.00
1
1.60
33
0.03
14
2.90
5
1.30
19
2.60
45
1.85
54
1.00
11
0.40
28
4.88
38
0.74
25
0.00
1
0.00
1
0.00
1
0.00
1
0.07
38
0.01
9
ccs_robtwo views1.03
17
0.66
60
2.06
27
0.20
51
0.02
28
1.30
5
0.03
14
4.09
42
2.05
54
1.08
19
1.20
41
1.07
15
0.20
16
5.59
50
0.72
24
0.00
1
0.00
1
0.00
1
0.00
1
0.16
58
0.05
44
NLCA_NET_v2_RVCtwo views1.04
18
0.55
44
2.52
37
0.56
79
1.00
96
1.37
13
0.00
1
3.43
19
1.35
22
2.24
38
0.35
19
2.18
34
0.58
35
4.28
25
0.34
6
0.00
1
0.00
1
0.00
1
0.01
66
0.00
1
0.00
1
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 views1.04
18
0.57
49
2.62
40
0.50
77
0.84
93
1.36
12
0.01
7
3.40
17
1.29
18
2.28
41
0.37
21
2.21
35
0.60
36
4.43
30
0.38
10
0.00
1
0.00
1
0.00
1
0.01
66
0.01
6
0.00
1
iResNet_ROBtwo views1.05
20
0.76
67
2.06
27
0.07
28
0.00
1
1.35
10
0.03
14
6.80
96
2.57
69
0.79
12
1.20
41
0.94
10
0.12
8
3.81
17
0.41
11
0.00
1
0.00
1
0.00
1
0.00
1
0.08
40
0.01
9
iResNetv2_ROBtwo views1.08
21
1.12
92
4.68
76
1.09
101
0.00
1
1.38
15
0.05
23
3.83
32
1.11
9
0.80
13
1.15
38
1.02
12
0.12
8
4.52
31
0.45
14
0.00
1
0.00
1
0.00
1
0.00
1
0.33
89
0.03
36
TDLMtwo views1.08
21
0.49
38
1.43
13
0.27
59
0.00
1
1.56
29
2.12
80
3.10
10
1.75
48
1.36
25
0.93
34
1.06
14
0.49
32
6.21
58
0.70
23
0.00
1
0.00
1
0.00
1
0.00
1
0.16
58
0.02
25
FADNet-RVC-Resampletwo views1.08
21
1.00
85
5.32
80
0.10
39
0.14
55
1.53
27
0.25
52
3.62
26
1.18
13
0.51
7
0.19
9
1.04
13
0.28
21
5.20
41
0.44
12
0.10
90
0.16
95
0.09
91
0.06
84
0.24
76
0.21
84
CFNettwo views1.08
21
0.62
55
2.26
31
0.26
58
0.03
33
1.40
16
0.04
19
4.45
57
2.09
55
0.48
6
0.69
30
1.31
17
0.24
19
6.63
68
0.83
27
0.00
1
0.00
1
0.00
1
0.00
1
0.17
63
0.04
40
PSMNet_ROBtwo views1.09
25
0.69
63
2.92
43
0.03
19
0.07
41
1.68
36
0.16
38
3.44
20
1.30
19
0.96
17
0.32
14
2.81
44
0.63
38
4.23
23
2.48
55
0.00
1
0.00
1
0.00
1
0.00
1
0.08
40
0.04
40
NVstereo2Dtwo views1.09
25
0.17
10
4.40
73
0.00
1
0.00
1
1.59
32
0.18
43
4.29
51
1.40
28
0.85
14
0.36
20
2.30
37
0.62
37
5.26
44
0.35
7
0.00
1
0.00
1
0.00
1
0.00
1
0.01
6
0.01
9
GANetREF_RVCpermissivetwo views1.10
27
0.95
81
2.98
44
0.06
25
0.00
1
2.23
45
0.18
43
3.95
38
2.53
67
0.56
8
0.32
14
1.89
27
0.66
42
5.03
39
0.51
17
0.00
1
0.00
1
0.00
1
0.00
1
0.05
32
0.04
40
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
FADNet-RVCtwo views1.10
27
1.07
90
4.79
77
0.03
19
0.00
1
1.60
33
0.08
26
3.77
30
1.70
45
0.35
4
0.07
3
0.61
8
0.66
42
5.32
45
1.11
31
0.01
65
0.08
87
0.00
1
0.00
1
0.48
98
0.18
80
CVANet_RVCtwo views1.10
27
0.39
26
1.69
17
0.28
60
0.00
1
1.45
21
0.48
59
3.26
12
1.64
44
2.93
48
1.22
43
1.51
18
0.37
27
6.09
57
0.50
16
0.00
1
0.00
1
0.00
1
0.00
1
0.18
67
0.01
9
FADNet_RVCtwo views1.15
30
1.06
89
4.34
72
0.02
10
0.06
38
2.77
58
0.04
19
3.50
24
0.80
6
0.35
4
0.32
14
0.65
9
0.09
6
5.20
41
1.56
37
0.24
102
0.34
104
0.11
93
0.33
104
0.65
105
0.60
99
ETE_ROBtwo views1.16
31
0.52
42
1.71
18
0.01
5
0.02
28
1.79
40
0.22
48
4.10
43
1.32
21
4.68
74
1.18
40
2.81
44
0.08
5
3.56
13
1.12
32
0.00
1
0.00
1
0.00
1
0.00
1
0.01
6
0.06
51
XPNet_ROBtwo views1.17
32
0.41
29
1.68
16
0.04
23
0.01
21
1.65
35
0.17
40
3.46
22
1.39
25
3.89
64
1.01
35
1.60
21
0.53
33
4.35
29
3.07
62
0.00
1
0.00
1
0.00
1
0.00
1
0.03
17
0.02
25
DLCB_ROBtwo views1.17
32
0.22
14
1.28
8
0.08
32
0.00
1
1.51
26
0.23
49
3.39
16
1.52
40
3.43
60
2.00
57
3.40
54
0.31
24
4.29
26
1.82
44
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
0.00
1
FADNettwo views1.20
34
0.93
80
4.51
74
0.03
19
0.06
38
2.56
54
0.16
38
4.02
40
1.42
30
0.32
2
0.06
2
1.66
23
1.07
46
5.56
49
0.87
28
0.19
96
0.02
68
0.01
72
0.01
66
0.50
100
0.03
36
NOSS_ROBtwo views1.29
35
0.22
14
1.35
12
0.64
84
0.16
60
2.56
54
0.00
1
4.14
45
2.37
62
2.29
42
1.76
53
2.84
46
0.02
3
6.80
73
0.45
14
0.00
1
0.00
1
0.00
1
0.00
1
0.04
27
0.11
65
NCCL2two views1.30
36
0.63
56
2.31
32
0.02
10
0.03
33
1.57
30
3.41
89
3.91
37
1.39
25
3.14
54
0.34
17
3.90
58
0.30
22
3.75
15
1.19
33
0.00
1
0.00
1
0.02
82
0.00
1
0.03
17
0.05
44
PA-Nettwo views1.31
37
0.59
52
3.65
60
0.06
25
0.15
58
1.44
20
0.54
62
4.08
41
2.71
71
0.78
11
1.16
39
2.32
39
0.64
39
5.96
54
1.92
47
0.01
65
0.00
1
0.00
1
0.00
1
0.04
27
0.06
51
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
LALA_ROBtwo views1.33
38
0.59
52
1.74
19
0.02
10
0.06
38
3.25
63
0.50
61
3.45
21
1.39
25
3.70
62
0.59
28
4.59
66
0.25
20
3.58
14
2.79
59
0.00
1
0.00
1
0.00
1
0.00
1
0.04
27
0.02
25
HSMtwo views1.36
39
0.40
27
1.07
4
0.01
5
0.30
74
1.73
39
0.03
14
5.89
84
1.22
14
1.83
29
1.36
47
4.78
69
1.30
51
6.61
67
0.57
20
0.00
1
0.00
1
0.00
1
0.00
1
0.02
15
0.00
1
HSM-Net_RVCpermissivetwo views1.40
40
0.07
1
1.58
14
0.00
1
0.01
21
1.88
42
0.05
23
7.24
97
1.71
47
3.63
61
3.63
67
3.21
52
0.36
25
4.25
24
0.44
12
0.01
65
0.01
58
0.00
1
0.00
1
0.01
6
0.00
1
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
StereoDRNettwo views1.48
41
0.75
64
4.58
75
0.05
24
0.10
51
2.49
52
1.10
70
4.62
61
1.47
37
3.38
59
0.62
29
2.29
36
0.65
41
4.52
31
2.89
60
0.00
1
0.03
71
0.00
1
0.01
66
0.05
32
0.07
57
RYNettwo views1.52
42
0.35
23
3.38
53
0.25
55
0.04
36
2.31
49
0.03
14
4.62
61
1.47
37
1.14
21
0.34
17
4.52
64
1.38
52
6.36
62
4.23
72
0.00
1
0.00
1
0.00
1
0.00
1
0.01
6
0.02
25
RASNettwo views1.53
43
0.31
20
1.97
23
0.09
37
3.20
108
1.69
37
0.06
25
3.27
13
2.12
56
0.77
10
1.29
45
5.32
77
1.69
59
6.60
66
2.17
53
0.00
1
0.00
1
0.00
1
0.00
1
0.02
15
0.02
25
CBMVpermissivetwo views1.56
44
0.24
19
1.92
22
0.31
63
0.03
33
2.90
60
4.33
95
3.79
31
1.81
49
4.98
75
1.73
51
2.70
42
1.68
58
3.81
17
0.92
29
0.00
1
0.00
1
0.00
1
0.00
1
0.05
32
0.08
59
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
PASMtwo views1.58
45
1.05
88
7.51
91
0.56
79
0.09
45
1.37
13
0.28
53
2.85
4
1.44
33
3.91
65
0.44
26
3.02
49
0.64
39
6.38
63
1.59
38
0.00
1
0.01
58
0.00
1
0.03
76
0.26
79
0.15
74
DRN-Testtwo views1.62
46
0.23
17
3.72
61
0.18
46
0.13
54
3.34
65
0.19
47
5.93
85
2.01
53
4.14
68
0.76
31
2.12
33
1.78
61
5.22
43
2.62
57
0.00
1
0.00
1
0.01
72
0.01
66
0.03
17
0.01
9
AANet_RVCtwo views1.66
47
0.88
77
3.52
57
0.06
25
0.00
1
1.22
2
0.00
1
3.25
11
2.51
66
2.21
37
4.49
71
3.65
55
0.20
16
7.10
80
3.42
66
0.50
107
0.04
76
0.00
1
0.00
1
0.08
40
0.05
44
Abc-Nettwo views1.67
48
0.56
47
3.10
48
0.08
32
0.22
63
2.24
46
0.13
35
5.05
74
1.42
30
2.27
39
2.72
58
4.29
59
3.53
82
5.76
52
2.01
49
0.04
81
0.00
1
0.00
1
0.00
1
0.03
17
0.02
25
NCC-stereotwo views1.67
48
0.56
47
3.10
48
0.08
32
0.22
63
2.24
46
0.13
35
5.05
74
1.42
30
2.27
39
2.72
58
4.29
59
3.53
82
5.76
52
2.01
49
0.04
81
0.00
1
0.00
1
0.00
1
0.03
17
0.02
25
CBMV_ROBtwo views1.70
50
0.11
7
1.98
24
0.46
74
0.00
1
2.85
59
0.91
66
4.27
49
2.21
60
4.02
67
5.00
75
2.97
47
2.32
69
5.40
46
1.35
35
0.00
1
0.00
1
0.00
1
0.00
1
0.03
17
0.06
51
DANettwo views1.74
51
0.22
14
3.06
46
0.75
88
0.44
81
2.46
51
0.35
56
3.40
17
1.41
29
4.22
69
1.56
48
5.03
73
0.56
34
6.29
60
5.00
80
0.01
65
0.00
1
0.00
1
0.01
66
0.04
27
0.08
59
RPtwo views1.76
52
0.55
44
2.59
39
0.22
53
0.60
91
1.70
38
0.98
68
4.31
53
1.99
52
3.12
53
5.04
76
4.41
62
2.96
74
4.72
34
1.90
45
0.01
65
0.00
1
0.00
1
0.01
66
0.14
55
0.01
9
Anonymous Stereotwo views1.78
53
1.37
97
7.79
92
0.42
71
0.07
41
1.49
25
3.20
87
3.31
14
1.84
50
2.75
46
1.14
37
1.84
25
0.09
6
6.89
76
3.32
65
0.00
1
0.00
1
0.00
1
0.00
1
0.09
48
0.01
9
SGM-Foresttwo views1.84
54
0.08
3
1.31
9
0.15
43
0.22
63
3.52
66
2.46
83
3.85
34
2.16
57
5.29
76
3.90
68
3.71
56
1.56
55
6.28
59
1.90
45
0.01
65
0.02
68
0.00
1
0.00
1
0.03
17
0.30
89
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
RGCtwo views1.86
55
0.67
61
3.45
56
0.07
28
0.26
68
3.29
64
0.11
34
4.78
68
1.46
36
3.26
55
4.51
72
5.14
74
2.68
72
5.46
47
1.94
48
0.00
1
0.00
1
0.00
1
0.00
1
0.08
40
0.09
62
GANettwo views1.90
56
0.43
30
1.83
21
0.39
69
0.00
1
2.24
46
3.37
88
3.90
36
1.44
33
2.17
36
9.66
87
2.57
41
1.09
47
6.88
75
1.81
43
0.00
1
0.00
1
0.02
82
0.00
1
0.10
49
0.03
36
DISCOtwo views1.93
57
0.18
12
2.44
34
0.67
85
0.53
87
5.25
79
0.00
1
5.43
78
1.56
42
2.09
35
0.91
33
6.48
83
0.89
45
8.55
91
3.60
69
0.00
1
0.00
1
0.00
1
0.00
1
0.06
36
0.01
9
edge stereotwo views1.94
58
0.40
27
3.82
65
0.15
43
0.09
45
1.81
41
0.45
58
4.77
66
2.21
60
3.36
57
3.94
69
8.39
90
1.72
60
5.75
51
1.69
41
0.00
1
0.00
1
0.01
72
0.00
1
0.17
63
0.01
9
Nwc_Nettwo views1.94
58
0.55
44
3.79
63
0.07
28
0.08
44
2.73
56
0.04
19
6.75
95
1.11
9
2.03
34
7.92
81
4.82
70
2.11
67
4.78
35
2.01
49
0.00
1
0.00
1
0.00
1
0.00
1
0.05
32
0.01
9
MFMNet_retwo views2.02
60
0.98
82
4.27
70
2.74
108
0.88
94
1.47
23
0.15
37
4.21
46
3.18
75
6.63
80
6.93
80
1.87
26
3.63
85
2.13
9
0.99
30
0.00
1
0.00
1
0.00
1
0.00
1
0.23
75
0.05
44
NaN_ROBtwo views2.02
60
0.58
50
2.65
41
0.29
61
0.28
73
3.19
62
5.19
100
3.84
33
2.55
68
5.31
77
1.73
51
2.30
37
1.96
65
6.78
72
3.09
63
0.01
65
0.06
83
0.03
84
0.10
91
0.10
49
0.37
94
stereogantwo views2.13
62
0.31
20
3.75
62
0.01
5
0.09
45
7.44
92
0.60
63
4.27
49
3.29
76
3.91
65
4.62
73
6.41
82
1.84
63
5.17
40
0.58
21
0.00
1
0.01
58
0.01
72
0.00
1
0.24
76
0.05
44
AF-Nettwo views2.17
63
0.65
58
3.53
58
0.00
1
0.09
45
2.02
44
0.48
59
6.44
88
1.70
45
3.36
57
10.22
88
4.95
72
2.81
73
4.34
28
2.66
58
0.00
1
0.00
1
0.00
1
0.00
1
0.13
54
0.01
9
PDISCO_ROBtwo views2.21
64
0.78
70
3.43
55
0.96
96
2.04
107
7.04
89
0.09
28
7.36
99
5.72
91
1.21
23
1.67
49
3.16
51
1.28
50
7.52
85
1.37
36
0.00
1
0.00
1
0.00
1
0.00
1
0.34
90
0.13
73
ADCReftwo views2.29
65
0.65
58
5.67
85
0.10
39
0.09
45
4.81
75
0.76
64
4.75
65
1.37
23
2.88
47
1.67
49
1.62
22
2.30
68
3.77
16
14.93
99
0.02
76
0.00
1
0.01
72
0.10
91
0.20
72
0.11
65
FBW_ROBtwo views2.36
66
0.36
24
3.16
50
0.37
68
0.14
55
4.42
74
0.18
43
6.64
92
3.05
74
2.98
51
2.85
60
4.86
71
1.11
48
11.00
102
4.53
73
0.16
95
0.05
80
0.37
105
0.09
88
0.19
71
0.78
104
DeepPrunerFtwo views2.50
67
0.87
76
14.90
102
0.17
45
0.31
75
1.41
17
0.29
54
5.62
80
9.97
97
1.02
18
0.43
25
2.01
28
1.65
57
6.85
74
3.99
70
0.00
1
0.01
58
0.07
88
0.10
91
0.24
76
0.12
72
XQCtwo views2.57
68
1.20
93
6.33
89
0.18
46
0.01
21
3.84
69
0.33
55
5.24
76
3.83
82
4.43
71
1.94
55
4.58
65
3.19
78
7.19
81
8.67
90
0.00
1
0.03
71
0.00
1
0.03
76
0.27
82
0.19
81
RTSCtwo views2.66
69
1.24
94
6.04
86
0.18
46
0.02
28
4.40
73
0.08
26
4.29
51
4.98
89
6.00
78
3.15
62
2.97
47
1.56
55
6.71
70
10.89
95
0.00
1
0.07
85
0.00
1
0.05
80
0.40
95
0.24
88
CSANtwo views2.70
70
0.75
64
3.41
54
0.18
46
0.12
52
4.18
71
4.05
92
4.11
44
4.36
86
6.03
79
6.24
79
4.60
67
4.74
89
7.29
82
3.43
67
0.02
76
0.03
71
0.01
72
0.02
75
0.18
67
0.22
85
ADCP+two views2.87
71
0.98
82
8.75
95
0.07
28
0.16
60
5.93
84
2.90
86
4.00
39
1.84
50
1.52
26
0.27
12
3.81
57
3.22
80
6.05
56
17.38
101
0.00
1
0.00
1
0.00
1
0.00
1
0.14
55
0.30
89
PWC_ROBbinarytwo views2.90
72
1.01
86
6.12
87
0.53
78
0.57
90
1.30
5
0.40
57
3.85
34
4.42
87
7.37
84
15.40
97
3.03
50
1.55
54
6.31
61
5.74
86
0.00
1
0.00
1
0.00
1
0.00
1
0.46
96
0.05
44
SHDtwo views3.01
73
1.02
87
5.52
83
0.45
73
0.39
78
4.12
70
0.17
40
6.68
93
12.04
100
7.20
83
3.36
64
4.40
61
2.58
71
4.69
33
7.17
88
0.01
65
0.04
76
0.00
1
0.03
76
0.22
74
0.17
78
PWCDC_ROBbinarytwo views3.06
74
1.38
98
3.81
64
0.14
42
0.00
1
3.14
61
0.02
11
4.40
55
12.35
101
1.08
19
23.69
105
2.01
28
2.01
66
3.87
20
2.43
54
0.24
102
0.00
1
0.00
1
0.00
1
0.61
104
0.08
59
SPS-STEREOcopylefttwo views3.08
75
0.49
38
2.49
36
0.21
52
0.24
66
3.67
68
1.24
71
5.80
83
2.43
65
10.68
95
8.00
82
9.48
94
3.59
84
8.11
88
5.08
81
0.01
65
0.01
58
0.00
1
0.00
1
0.04
27
0.11
65
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
NVStereoNet_ROBtwo views3.25
76
0.46
33
2.71
42
0.34
66
0.56
89
2.36
50
0.87
65
4.56
59
4.47
88
4.55
72
14.09
96
11.93
96
3.23
81
9.79
98
3.14
64
0.43
105
0.03
71
0.53
107
0.31
103
0.36
93
0.38
95
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
ADCLtwo views3.30
77
0.81
73
6.76
90
0.25
55
0.25
67
6.93
87
4.65
96
4.79
69
2.62
70
4.61
73
3.36
64
5.22
75
4.55
88
5.54
48
14.91
98
0.05
84
0.01
58
0.11
93
0.13
94
0.28
83
0.10
63
MFN_U_SF_DS_RVCtwo views3.30
77
2.61
104
7.84
93
0.09
37
0.26
68
12.92
101
7.62
106
4.25
47
3.64
80
2.95
49
3.99
70
4.61
68
1.22
49
8.36
89
1.77
42
0.59
108
1.15
110
0.04
85
0.85
108
0.48
98
0.67
102
SAMSARAtwo views3.32
79
1.07
90
6.25
88
0.86
93
0.31
75
7.55
93
3.87
91
5.26
77
4.20
85
7.43
85
3.35
63
10.04
95
3.83
87
6.98
79
4.83
77
0.00
1
0.14
92
0.00
1
0.05
80
0.17
63
0.22
85
DPSNettwo views3.38
80
0.60
54
9.55
96
0.61
82
0.46
82
5.10
78
1.44
75
11.17
106
3.92
83
2.57
44
0.88
32
5.90
78
9.68
103
7.63
86
7.25
89
0.08
86
0.16
95
0.01
72
0.07
85
0.34
90
0.17
78
STTStereo_v2two views3.49
81
0.47
35
4.03
68
1.61
102
1.91
103
10.78
99
1.92
77
3.06
8
1.26
15
8.59
89
12.06
90
6.13
80
7.82
97
4.78
35
4.77
75
0.08
86
0.12
90
0.07
88
0.15
95
0.16
58
0.11
65
G-Nettwo views3.49
81
0.47
35
4.03
68
1.61
102
1.91
103
10.78
99
1.92
77
3.06
8
1.26
15
8.59
89
12.06
90
6.13
80
7.82
97
4.78
35
4.77
75
0.08
86
0.12
90
0.07
88
0.15
95
0.16
58
0.11
65
MDST_ROBtwo views3.50
83
0.07
1
3.94
66
2.66
107
1.46
101
12.94
102
0.97
67
7.32
98
2.42
64
14.82
103
8.08
83
2.74
43
1.51
53
8.44
90
2.48
55
0.00
1
0.00
1
0.00
1
0.00
1
0.01
6
0.11
65
pmcnntwo views3.61
84
0.68
62
5.38
82
0.61
82
1.46
101
2.75
57
1.26
72
5.04
73
1.57
43
8.77
91
11.67
89
18.69
105
2.38
70
6.03
55
5.61
85
0.00
1
0.07
85
0.00
1
0.00
1
0.21
73
0.07
57
ADCPNettwo views3.63
85
0.82
74
13.20
100
0.08
32
0.43
80
7.02
88
2.23
81
4.77
66
2.41
63
4.31
70
1.96
56
7.57
88
5.16
90
6.45
65
13.80
97
0.02
76
0.75
108
0.04
85
0.90
109
0.16
58
0.53
98
SuperBtwo views3.70
86
1.77
100
15.73
103
0.31
63
0.05
37
4.37
72
1.67
76
3.58
25
11.57
98
2.36
43
1.32
46
2.03
30
1.91
64
6.95
78
9.01
91
0.05
84
0.03
71
0.10
92
0.08
86
11.00
110
0.10
63
AnyNet_C32two views3.93
87
2.58
103
12.07
99
0.39
69
0.49
85
5.50
81
6.45
104
4.44
56
2.19
59
6.86
82
2.94
61
5.28
76
3.20
79
6.63
68
18.37
103
0.08
86
0.04
76
0.20
99
0.09
88
0.57
103
0.19
81
ADCMidtwo views3.95
88
1.34
95
10.91
98
0.18
46
0.38
77
5.02
77
1.27
73
4.87
70
2.94
73
11.15
96
3.60
66
6.76
85
5.21
92
6.40
64
16.96
100
0.12
92
0.08
87
0.55
108
0.46
105
0.53
102
0.19
81
SGM_RVCbinarytwo views4.07
89
0.38
25
1.98
24
0.84
89
0.26
68
6.20
85
2.37
82
6.45
89
3.57
78
11.56
98
9.27
86
15.55
102
6.78
94
10.07
99
4.60
74
0.22
97
0.25
100
0.20
99
0.23
100
0.26
79
0.34
92
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
SGM+DAISYtwo views4.39
90
1.34
95
5.18
79
1.01
98
0.88
94
5.61
82
2.86
85
4.88
71
3.53
77
13.29
100
13.95
95
12.66
98
7.55
95
7.98
87
5.39
84
0.23
101
0.18
98
0.15
97
0.24
101
0.29
84
0.64
100
WCMA_ROBtwo views4.54
91
0.48
37
3.19
51
0.49
76
0.63
92
5.38
80
2.57
84
4.70
63
3.77
81
14.17
102
19.02
100
15.47
101
9.03
100
6.73
71
4.85
78
0.01
65
0.05
80
0.01
72
0.01
66
0.08
40
0.16
75
SANettwo views4.63
92
0.89
78
5.34
81
0.30
62
0.09
45
7.43
91
3.75
90
6.53
90
16.16
104
6.79
81
13.29
93
12.44
97
7.74
96
7.38
83
4.11
71
0.00
1
0.01
58
0.00
1
0.00
1
0.10
49
0.31
91
MSMD_ROBtwo views4.94
93
0.43
30
2.58
38
0.12
41
0.01
21
8.43
96
1.09
69
4.25
47
4.11
84
13.11
99
26.36
107
13.95
100
13.88
106
7.45
84
2.92
61
0.10
90
0.01
58
0.00
1
0.00
1
0.01
6
0.02
25
ADCStwo views5.09
94
2.15
101
16.58
104
0.34
66
0.07
41
5.91
83
4.26
94
6.73
94
6.99
94
9.44
93
6.09
78
6.57
84
5.17
91
9.54
96
21.38
108
0.00
1
0.00
1
0.00
1
0.00
1
0.36
93
0.23
87
MSC_U_SF_DS_RVCtwo views5.12
95
2.33
102
8.42
94
0.33
65
0.27
71
24.60
111
8.84
108
5.50
79
9.75
96
3.82
63
8.16
84
8.53
91
1.82
62
9.44
95
5.18
82
1.02
109
1.65
111
0.33
104
0.47
106
1.06
107
0.90
106
MeshStereopermissivetwo views5.78
96
0.75
64
3.04
45
0.25
55
0.14
55
7.88
95
2.09
79
9.64
102
5.18
90
20.48
105
19.06
101
22.80
106
9.54
102
9.64
97
4.88
79
0.00
1
0.00
1
0.00
1
0.00
1
0.29
84
0.03
36
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
PVDtwo views5.81
97
0.99
84
5.56
84
0.72
87
0.51
86
6.72
86
0.18
43
9.90
104
23.69
108
11.45
97
19.26
102
8.18
89
9.46
101
9.24
93
9.49
92
0.02
76
0.16
95
0.00
1
0.08
86
0.18
67
0.40
96
FC-DCNNcopylefttwo views6.09
98
0.18
12
2.35
33
0.42
71
0.53
87
7.80
94
1.42
74
8.14
100
7.07
95
19.05
104
22.10
104
22.85
107
13.84
105
9.42
94
6.44
87
0.00
1
0.01
58
0.00
1
0.00
1
0.01
6
0.06
51
AnyNet_C01two views6.59
99
4.31
107
36.24
108
0.96
96
0.46
82
9.96
97
5.74
101
6.02
87
3.63
79
7.59
86
8.56
85
8.83
92
3.14
77
13.74
104
20.83
107
0.12
92
0.08
87
0.25
102
0.09
88
1.01
106
0.34
92
LSMtwo views7.38
100
1.68
99
22.24
106
3.99
109
40.72
112
3.54
67
4.79
99
4.74
64
6.65
92
10.25
94
13.51
94
4.49
63
3.81
86
6.91
77
5.31
83
0.00
1
0.04
76
0.00
1
0.00
1
0.26
79
14.60
111
ELAS_RVCcopylefttwo views7.69
101
0.80
71
4.90
78
1.06
99
1.02
97
10.00
98
9.43
109
9.75
103
14.69
103
21.83
106
20.27
103
17.88
103
16.41
108
13.01
103
10.44
94
0.22
97
0.55
106
0.28
103
0.19
97
0.35
92
0.65
101
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
DispFullNettwo views7.91
102
23.96
112
13.67
101
14.71
111
7.88
110
4.83
76
0.04
19
5.01
72
2.89
72
8.86
92
4.96
74
6.11
79
10.72
104
10.18
100
3.51
68
2.44
110
0.33
103
15.15
111
4.65
111
12.33
111
5.95
110
ELAScopylefttwo views8.05
103
0.76
67
4.02
67
0.95
95
1.02
97
14.85
105
6.99
105
12.80
108
11.73
99
22.43
107
27.75
108
17.88
103
15.19
107
10.96
101
11.31
96
0.22
97
0.56
107
0.23
101
0.22
99
0.50
100
0.71
103
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
PWCKtwo views8.72
104
5.37
108
16.61
105
5.74
110
0.12
52
19.88
107
11.41
110
11.31
107
13.42
102
14.07
101
12.84
92
13.20
99
8.05
99
15.44
105
10.25
93
5.95
111
0.22
99
2.09
110
0.05
80
7.50
109
0.92
107
RTStwo views9.75
105
2.69
105
65.25
111
0.85
90
1.99
105
14.36
103
4.75
97
5.69
81
17.29
105
7.97
87
19.00
98
6.83
86
3.02
75
26.15
109
18.62
105
0.00
1
0.15
93
0.00
1
0.00
1
0.29
84
0.16
75
RTSAtwo views9.75
105
2.69
105
65.25
111
0.85
90
1.99
105
14.36
103
4.75
97
5.69
81
17.29
105
7.97
87
19.00
98
6.83
86
3.02
75
26.15
109
18.62
105
0.00
1
0.15
93
0.00
1
0.00
1
0.29
84
0.16
75
MADNet+two views10.71
107
14.57
110
70.07
113
0.58
81
0.47
84
19.89
108
4.20
93
11.05
105
6.87
93
3.03
52
5.99
77
9.01
93
6.35
93
31.95
112
28.31
111
0.22
97
0.26
102
0.01
72
0.04
79
0.46
96
0.87
105
SGM-ForestMtwo views12.56
108
0.80
71
4.31
71
1.07
100
0.41
79
21.56
109
8.57
107
17.42
109
19.67
107
29.93
108
28.49
110
45.83
111
27.22
110
26.05
108
18.40
104
0.14
94
0.25
100
0.18
98
0.28
102
0.18
67
0.42
97
LE_ROBtwo views12.79
109
0.64
57
10.26
97
2.13
105
1.29
99
7.25
90
5.81
102
9.09
101
49.27
112
57.55
112
24.58
106
23.68
108
31.13
112
9.00
92
23.57
110
0.04
81
0.06
83
0.13
96
0.20
98
0.08
40
0.11
65
MANEtwo views14.02
110
0.85
75
3.53
58
2.13
105
3.75
109
18.44
106
6.14
103
22.67
111
27.90
110
33.01
110
41.08
112
41.26
110
30.90
111
23.39
106
17.99
102
0.47
106
0.76
109
0.91
109
4.04
110
0.17
63
1.00
108
BEATNet-Init1two views17.91
111
7.45
109
36.82
109
1.88
104
1.38
100
37.29
112
14.71
111
21.90
110
24.20
109
38.07
111
36.59
111
50.41
112
32.00
113
29.42
111
21.83
109
0.33
104
0.39
105
0.47
106
0.54
107
1.50
108
1.07
109
DPSimNet_ROBtwo views25.03
112
19.56
111
24.93
107
24.36
112
17.85
111
22.25
110
25.93
112
28.52
112
29.10
111
32.61
109
28.25
109
32.87
109
26.42
109
24.51
107
40.42
112
16.75
112
13.93
112
25.19
112
22.38
112
19.96
112
24.75
112
MADNet++two views47.78
113
34.15
113
37.11
110
42.94
113
41.23
113
64.94
113
30.63
113
60.79
113
53.26
113
72.08
113
61.65
113
57.68
113
54.96
114
58.03
113
59.24
113
43.01
113
36.17
113
36.07
113
22.50
113
47.50
113
41.67
113
MEDIAN_ROBtwo views96.83
114
99.41
114
98.66
115
94.75
114
94.23
114
93.08
114
90.54
114
95.61
114
94.75
114
97.65
116
97.73
114
98.30
114
96.57
115
94.51
114
93.74
114
99.78
119
98.24
114
99.99
116
99.89
114
99.48
116
99.67
116
DPSMtwo views99.17
115
100.00
118
100.00
118
96.28
115
98.84
115
100.00
116
100.00
116
100.00
117
100.00
115
100.00
117
100.00
115
100.00
117
100.00
116
100.00
118
100.00
118
93.66
114
100.00
115
100.00
117
100.00
115
96.31
114
98.27
114
DPSM_ROBtwo views99.17
115
100.00
118
100.00
118
96.28
115
98.84
115
100.00
116
100.00
116
100.00
117
100.00
115
100.00
117
100.00
115
100.00
117
100.00
116
100.00
118
100.00
118
93.66
114
100.00
115
100.00
117
100.00
115
96.31
114
98.27
114
AVERAGE_ROBtwo views99.23
117
99.81
115
97.56
114
100.00
120
100.00
117
96.24
115
91.02
115
100.00
117
100.00
115
100.00
117
100.00
115
100.00
117
100.00
116
99.98
117
99.97
115
100.00
120
100.00
115
100.00
117
100.00
115
100.00
120
100.00
119
DGTPSM_ROBtwo views99.45
118
99.94
116
99.98
116
96.58
117
100.00
117
100.00
116
100.00
116
99.81
115
100.00
115
95.86
114
100.00
115
99.32
115
100.00
116
99.79
115
99.99
116
98.34
116
100.00
115
99.86
114
100.00
115
99.52
117
99.99
118
DPSMNet_ROBtwo views99.46
119
99.94
116
99.98
116
96.59
118
100.00
117
100.00
116
100.00
116
99.81
115
100.00
115
95.87
115
100.00
115
99.32
115
100.00
116
99.79
115
99.99
116
98.46
117
100.00
115
99.86
114
100.00
115
99.52
117
100.00
119
LSM0two views99.94
120
100.00
118
100.00
118
99.84
119
100.00
117
100.00
116
100.00
116
100.00
117
100.00
115
100.00
117
100.00
115
100.00
117
100.00
116
100.00
118
100.00
118
99.22
118
100.00
115
100.00
117
100.00
115
99.99
119
99.83
117
MSMDNettwo views0.42
29