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
sort bysort bysort bysort bysort bysort bysort bysorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
DN-CSS_ROBtwo views2.69
3
1.40
45
5.34
29
2.31
39
0.75
16
3.14
4
0.06
1
6.11
1
3.87
3
5.34
5
12.18
29
2.34
3
1.22
3
7.84
12
1.48
3
0.03
30
0.00
1
0.00
1
0.00
1
0.35
49
0.03
7
R-Stereotwo views2.44
1
0.32
1
1.93
1
0.94
2
0.16
2
3.67
6
0.61
12
6.37
2
3.08
1
9.14
25
17.44
51
1.80
1
0.77
1
1.76
1
0.70
1
0.00
1
0.01
22
0.00
1
0.00
1
0.01
1
0.03
7
R-Stereo Traintwo views2.44
1
0.32
1
1.93
1
0.94
2
0.16
2
3.67
6
0.61
12
6.37
2
3.08
1
9.14
25
17.44
51
1.80
1
0.77
1
1.76
1
0.70
1
0.00
1
0.01
22
0.00
1
0.00
1
0.01
1
0.03
7
MLCVtwo views3.44
9
0.88
17
5.60
31
1.39
9
0.25
4
4.36
13
0.33
5
7.25
4
7.28
9
9.17
27
12.24
30
5.09
6
2.47
10
9.15
31
3.23
14
0.00
1
0.00
1
0.00
1
0.00
1
0.10
19
0.02
2
iResNettwo views3.68
12
0.91
20
7.94
51
2.97
55
0.34
6
4.44
17
0.48
10
7.70
5
9.74
26
7.72
16
12.74
33
4.03
5
2.87
13
8.05
16
3.37
16
0.02
21
0.01
22
0.00
1
0.00
1
0.10
19
0.09
19
CFNet_RVCtwo views3.31
7
0.94
23
2.69
4
1.50
12
2.38
50
2.81
2
0.68
15
8.35
6
7.43
10
4.45
1
9.94
14
10.20
28
4.60
25
6.49
5
3.41
17
0.00
1
0.00
1
0.03
54
0.00
1
0.22
40
0.03
7
ccstwo views3.37
8
1.16
36
3.89
16
2.94
54
0.78
19
4.78
18
0.33
5
9.00
7
7.77
14
5.90
6
10.84
19
7.74
18
2.31
9
7.76
11
1.98
6
0.00
1
0.00
1
0.00
1
0.00
1
0.16
30
0.06
13
Anonymous Stereotwo views6.16
38
3.15
73
23.75
81
2.97
55
2.48
53
4.39
16
13.30
78
9.21
8
9.86
27
9.56
28
8.76
10
6.79
11
1.99
7
13.50
50
13.04
61
0.01
16
0.05
40
0.00
1
0.06
54
0.22
40
0.19
36
HITNettwo views2.79
4
0.77
13
4.02
17
2.03
30
0.11
1
5.58
22
0.59
11
9.24
9
5.15
5
6.42
10
7.26
4
3.66
4
2.92
14
4.07
3
3.87
22
0.00
1
0.00
1
0.00
1
0.00
1
0.06
15
0.02
2
DANettwo views6.02
36
1.23
39
8.45
53
3.86
68
3.94
67
7.64
31
1.34
24
9.51
10
7.00
7
13.39
51
15.53
46
15.99
50
7.02
42
12.14
43
12.37
58
0.19
51
0.12
55
0.02
48
0.03
45
0.13
27
0.56
58
DLCB_ROBtwo views4.51
24
0.91
20
3.78
14
2.19
35
1.07
29
6.28
24
3.09
29
9.78
11
7.72
12
10.65
34
12.97
34
13.91
44
3.71
17
8.72
26
5.30
28
0.00
1
0.00
1
0.00
1
0.00
1
0.03
9
0.10
25
StereoDRNet-Refinedtwo views4.46
23
0.62
12
3.80
15
1.92
25
0.40
8
9.35
35
0.15
2
10.02
12
8.83
21
12.69
45
11.62
26
9.34
24
3.87
18
8.06
17
8.02
37
0.00
1
0.00
1
0.01
39
0.05
53
0.20
35
0.26
43
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
AANet_RVCtwo views5.01
28
1.74
52
6.38
41
1.96
29
1.29
33
2.26
1
1.69
26
10.07
13
18.53
57
7.88
18
18.15
53
8.49
23
2.70
12
10.59
39
7.04
34
0.96
80
0.15
61
0.02
48
0.00
1
0.13
27
0.12
27
CBMV_ROBtwo views4.14
18
0.52
7
3.14
8
1.30
7
0.77
18
6.92
28
1.97
28
10.11
14
9.58
24
8.92
24
14.20
41
7.12
15
5.90
36
8.65
25
3.50
20
0.01
16
0.05
40
0.00
1
0.00
1
0.04
10
0.09
19
ccs_robtwo views3.63
11
1.12
34
4.42
21
2.52
43
0.91
23
5.50
21
0.21
3
10.11
14
9.11
22
6.55
12
11.28
24
8.32
22
2.55
11
7.66
9
2.01
7
0.00
1
0.00
1
0.00
1
0.00
1
0.20
35
0.08
17
NOSS_ROBtwo views3.30
6
0.46
6
2.62
3
2.08
31
1.01
27
5.60
23
0.74
17
10.37
16
11.48
37
5.15
4
8.43
9
5.67
7
1.73
6
7.97
14
2.34
8
0.02
21
0.06
45
0.00
1
0.00
1
0.07
16
0.14
31
CC-Net-ROBtwo views3.84
14
1.07
28
5.23
27
2.65
47
2.96
56
4.22
12
0.69
16
10.43
17
7.72
12
8.78
22
8.29
8
9.61
26
4.02
21
7.16
8
3.65
21
0.13
46
0.03
33
0.02
48
0.03
45
0.05
12
0.03
7
NVstereo2Dtwo views4.51
24
0.82
15
6.86
44
3.28
60
3.38
61
8.16
32
3.13
30
10.51
18
15.15
46
4.90
3
6.89
2
7.87
19
4.78
29
9.88
35
3.91
23
0.01
16
0.00
1
0.00
1
0.06
54
0.02
4
0.58
59
PASMtwo views7.90
49
4.22
77
21.97
78
3.25
59
3.29
59
5.39
20
6.57
55
10.57
19
19.09
59
12.77
47
13.92
39
18.11
56
9.51
53
13.79
53
10.77
54
0.19
51
0.45
77
0.29
71
1.08
82
1.49
78
1.19
73
TDLMtwo views4.11
17
1.11
33
3.54
11
1.62
16
1.04
28
3.91
9
7.41
61
10.60
20
10.67
32
6.38
9
12.59
32
5.95
9
4.77
28
8.79
28
3.04
12
0.58
73
0.00
1
0.01
39
0.00
1
0.19
34
0.12
27
NLCA_NET_v2_RVCtwo views3.84
14
1.06
27
5.23
27
2.72
49
3.27
58
4.36
13
0.61
12
10.71
21
7.56
11
8.75
21
7.89
6
9.86
27
3.90
19
7.15
7
3.44
18
0.14
47
0.02
29
0.02
48
0.03
45
0.04
10
0.03
7
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
ADCReftwo views7.27
45
1.38
44
16.37
70
2.52
43
3.30
60
11.63
49
3.16
31
10.80
22
9.35
23
13.03
50
25.27
63
8.17
21
8.92
50
8.06
17
21.81
75
0.15
48
0.08
50
0.16
68
0.34
72
0.38
52
0.58
59
CVANet_RVCtwo views4.16
19
1.16
36
3.60
12
1.94
28
1.46
34
3.92
10
4.68
43
10.89
23
8.34
19
7.58
15
10.84
19
10.27
29
6.62
40
8.56
24
2.69
9
0.39
65
0.00
1
0.00
1
0.01
33
0.21
39
0.09
19
CFNettwo views3.72
13
1.10
32
5.03
25
2.49
42
1.59
36
4.90
19
0.22
4
11.38
24
9.88
28
4.80
2
11.25
23
6.44
10
3.68
16
8.33
22
3.00
11
0.00
1
0.00
1
0.00
1
0.00
1
0.22
40
0.07
15
SGM-Foresttwo views4.96
26
0.32
1
2.84
6
1.21
5
0.64
12
10.23
44
6.64
56
11.55
25
10.98
33
10.94
37
13.59
37
11.65
34
4.30
24
8.94
30
4.63
26
0.11
43
0.04
37
0.00
1
0.00
1
0.05
12
0.46
53
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
WCMA_ROBtwo views9.21
57
0.87
16
7.37
47
2.54
45
2.13
46
13.59
57
5.80
47
11.64
26
14.01
42
24.43
76
32.99
76
27.09
75
18.02
66
12.51
47
9.85
50
0.81
77
0.07
48
0.01
39
0.01
33
0.16
30
0.23
38
HSM-Net_RVCpermissivetwo views4.20
20
0.32
1
2.76
5
0.63
1
0.69
14
6.95
29
1.69
26
11.96
27
8.36
20
8.83
23
12.17
28
15.18
49
4.21
23
6.91
6
3.30
15
0.02
21
0.02
29
0.00
1
0.00
1
0.01
1
0.01
1
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
GANettwo views6.22
39
1.07
28
4.07
19
2.27
37
0.89
21
9.19
34
9.52
66
12.02
28
8.13
17
10.72
35
29.09
69
13.86
42
7.52
45
11.00
40
4.39
24
0.36
64
0.00
1
0.02
48
0.02
38
0.12
25
0.08
17
NCCL2two views5.88
34
1.59
49
5.44
30
1.87
21
0.92
24
9.55
38
11.55
75
12.11
29
9.94
29
9.67
29
8.85
11
22.28
67
7.41
43
8.78
27
7.17
35
0.01
16
0.00
1
0.03
54
0.00
1
0.13
27
0.23
38
ADCPNettwo views9.54
59
2.39
63
31.46
83
2.09
32
1.60
37
16.71
66
6.39
54
12.11
29
11.45
36
13.53
52
21.45
57
19.41
60
10.94
58
14.38
56
21.54
73
0.27
59
1.16
81
0.39
77
1.49
85
0.58
59
1.45
77
DISCOtwo views6.28
40
0.57
9
5.78
34
3.43
62
1.17
30
11.22
47
3.39
34
12.14
31
16.16
50
6.52
11
11.22
22
16.96
52
6.32
37
19.51
71
10.74
52
0.00
1
0.00
1
0.00
1
0.00
1
0.35
49
0.11
26
AdaStereotwo views3.09
5
0.58
10
3.04
7
2.84
50
0.48
11
4.08
11
1.29
23
12.16
32
7.77
14
6.03
7
9.62
13
5.79
8
1.53
5
4.56
4
1.93
5
0.00
1
0.00
1
0.00
1
0.00
1
0.02
4
0.02
2
Xiao Song, Guorun Yang, Xinge Zhu, Hui Zhou, Zhe Wang, Jianping Shi: AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching. ArXiv
HSMtwo views4.00
16
0.79
14
3.16
9
1.59
15
2.17
48
6.77
27
1.11
19
12.28
33
6.35
6
6.75
13
8.11
7
13.90
43
5.37
33
8.85
29
2.71
10
0.00
1
0.00
1
0.00
1
0.00
1
0.02
4
0.02
2
PA-Nettwo views4.98
27
1.47
47
7.42
48
2.40
40
2.14
47
8.73
33
3.64
36
12.42
34
13.11
40
7.03
14
7.57
5
7.88
20
6.52
39
10.16
36
7.82
36
0.02
21
0.03
33
0.00
1
0.00
1
0.11
23
1.07
71
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
CBMVpermissivetwo views5.35
30
0.91
20
3.67
13
1.62
16
0.44
10
10.09
42
7.19
60
12.49
35
12.33
39
12.22
41
14.69
42
10.93
30
6.48
38
8.51
23
4.96
27
0.02
21
0.15
61
0.00
1
0.00
1
0.17
32
0.17
34
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
ETE_ROBtwo views5.80
32
1.77
54
6.33
40
1.44
11
0.78
19
6.43
26
6.90
57
12.53
36
8.08
16
12.93
49
14.89
43
21.13
66
5.87
35
9.83
34
6.57
32
0.04
33
0.01
22
0.00
1
0.02
38
0.08
18
0.33
46
pmcnntwo views7.72
48
1.27
41
9.42
56
2.91
51
3.14
57
9.44
36
6.23
50
12.56
37
16.51
52
14.53
55
24.08
60
27.44
76
8.49
48
9.32
32
8.44
41
0.06
38
0.08
50
0.00
1
0.00
1
0.30
45
0.15
32
ADCP+two views8.09
51
1.79
55
14.50
67
1.54
13
4.28
69
16.57
64
5.20
45
12.80
38
11.20
35
12.83
48
17.07
49
11.02
31
10.80
57
17.59
66
23.18
78
0.03
30
0.05
40
0.01
39
0.18
63
0.39
55
0.81
67
XPNet_ROBtwo views6.03
37
1.22
38
5.61
32
2.56
46
0.90
22
6.32
25
7.07
58
12.92
39
8.30
18
14.76
57
15.13
45
19.84
62
6.66
41
10.36
37
8.58
42
0.02
21
0.04
37
0.00
1
0.03
45
0.11
23
0.24
40
AnyNet_C32two views10.98
68
5.58
79
22.79
79
4.16
69
5.83
77
15.64
61
14.30
80
13.18
40
17.15
55
16.44
61
20.52
55
14.68
47
13.44
63
22.46
75
30.08
86
0.17
50
0.26
69
0.36
75
0.36
73
1.23
74
0.91
69
RYNettwo views6.34
41
0.89
19
5.88
35
1.41
10
4.48
74
15.97
62
4.18
39
13.41
41
16.49
51
10.81
36
7.00
3
14.33
46
8.72
49
9.43
33
13.71
62
0.00
1
0.01
22
0.00
1
0.00
1
0.02
4
0.07
15
DeepPrunerFtwo views6.75
44
2.69
66
23.31
80
3.68
64
7.16
81
3.78
8
4.29
40
13.42
42
20.13
63
8.13
20
10.46
15
7.18
16
8.06
46
11.10
41
9.44
47
0.24
54
0.15
61
0.29
71
0.42
75
0.66
63
0.45
51
MSMD_ROBtwo views9.28
58
1.09
31
4.65
23
1.58
14
0.39
7
16.52
63
4.41
42
13.60
43
14.87
45
22.34
69
39.89
83
25.67
72
20.71
76
12.42
46
6.98
33
0.34
63
0.03
33
0.00
1
0.00
1
0.05
12
0.09
19
iResNetv2_ROBtwo views4.28
22
1.43
46
7.17
46
2.91
51
1.26
31
4.36
13
1.62
25
13.64
44
10.25
31
9.83
30
11.41
25
7.68
17
4.00
20
7.75
10
1.85
4
0.00
1
0.00
1
0.00
1
0.00
1
0.37
51
0.09
19
DeepPruner_ROBtwo views3.52
10
1.14
35
4.06
18
1.12
4
1.65
39
3.65
5
0.83
18
13.96
45
4.47
4
7.80
17
10.84
19
7.05
14
2.16
8
8.14
20
3.08
13
0.07
39
0.03
33
0.00
1
0.01
33
0.32
46
0.06
13
ADCLtwo views10.16
64
2.11
60
19.36
76
1.92
25
1.88
45
22.23
75
8.91
62
14.04
46
23.56
68
14.62
56
26.19
64
12.75
38
13.59
64
16.06
63
22.95
77
0.26
56
0.18
64
0.75
80
0.65
77
0.69
64
0.58
59
ADCMidtwo views10.24
65
3.13
71
20.70
77
2.21
36
2.39
51
11.23
48
6.19
49
14.17
47
11.19
34
23.20
75
22.25
58
17.89
55
19.54
72
18.51
69
26.21
81
0.45
71
0.42
76
1.10
82
1.29
83
1.56
80
1.18
72
LE_ROBtwo views16.73
83
1.28
43
11.61
60
3.72
66
1.65
39
16.67
65
9.17
64
14.39
48
55.91
91
63.81
91
40.86
86
35.94
84
37.73
90
14.24
55
26.87
82
0.05
36
0.10
53
0.13
66
0.22
66
0.12
25
0.15
32
CSANtwo views7.62
46
1.60
50
6.56
42
1.83
19
0.66
13
12.40
52
10.52
73
14.45
49
21.32
65
14.19
53
15.98
48
17.84
54
13.02
62
12.32
45
8.38
40
0.09
41
0.07
48
0.03
54
0.04
51
0.33
47
0.67
64
PWCDC_ROBbinarytwo views7.92
50
3.17
74
7.48
49
5.73
80
4.40
70
10.45
45
0.35
7
14.52
50
28.19
74
10.36
33
31.27
72
7.04
13
9.14
52
13.22
49
8.78
44
2.74
84
0.02
29
0.00
1
0.00
1
1.31
76
0.17
34
XQCtwo views8.43
54
3.58
75
16.40
71
2.92
53
2.17
48
13.22
56
3.60
35
14.64
51
25.86
71
11.87
39
12.04
27
15.06
48
10.67
56
15.24
60
19.41
68
0.39
65
0.08
50
0.05
61
0.07
56
0.84
66
0.45
51
NaN_ROBtwo views6.00
35
1.24
40
6.29
39
1.34
8
1.68
41
9.60
40
10.31
71
15.09
52
15.79
48
12.62
44
8.95
12
11.67
35
5.83
34
11.78
42
6.41
31
0.05
36
0.13
57
0.08
63
0.20
65
0.22
40
0.79
66
GANetREF_RVCpermissivetwo views6.56
42
2.89
68
7.58
50
3.41
61
0.40
8
12.96
54
9.58
67
15.09
52
17.25
56
10.33
32
10.62
18
12.27
37
8.16
47
12.21
44
4.53
25
0.41
67
0.00
1
0.00
1
0.02
38
3.12
83
0.39
48
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
iResNet_ROBtwo views4.23
21
1.02
25
4.90
24
2.18
34
0.93
25
2.92
3
0.37
8
15.10
54
16.91
54
7.89
19
10.51
17
7.03
12
3.07
15
8.16
21
3.46
19
0.01
16
0.00
1
0.00
1
0.00
1
0.10
19
0.02
2
LSMtwo views14.01
74
5.95
80
33.49
84
6.78
82
43.61
91
10.22
43
9.98
70
15.16
55
22.93
67
23.07
74
32.34
75
18.52
57
12.67
59
15.45
61
11.10
55
0.16
49
0.51
78
0.09
65
0.32
70
1.08
71
16.85
90
RTSAtwo views18.87
86
9.32
85
86.48
91
4.95
77
6.10
78
42.08
88
14.70
81
15.49
56
41.06
86
22.65
71
32.32
73
13.77
40
19.54
72
37.98
87
28.96
83
0.41
67
0.23
67
0.00
1
0.02
38
0.91
67
0.50
55
RTStwo views18.87
86
9.32
85
86.48
91
4.95
77
6.10
78
42.08
88
14.70
81
15.49
56
41.06
86
22.65
71
32.32
73
13.77
40
19.54
72
37.98
87
28.96
83
0.41
67
0.23
67
0.00
1
0.02
38
0.91
67
0.50
55
PSMNet_ROBtwo views5.02
29
1.63
51
6.03
37
1.90
24
1.83
44
9.57
39
6.35
53
15.58
58
7.23
8
6.15
8
10.48
16
12.22
36
4.16
22
8.02
15
8.71
43
0.02
21
0.01
22
0.01
39
0.10
61
0.20
35
0.12
27
LALA_ROBtwo views6.58
43
1.80
56
6.25
38
1.26
6
0.94
26
10.08
41
9.02
63
16.00
59
11.51
38
12.74
46
13.02
35
24.77
70
5.25
31
10.56
38
8.02
37
0.04
33
0.05
40
0.00
1
0.02
38
0.10
19
0.25
41
DispFullNettwo views17.47
85
26.01
88
33.98
85
22.58
89
20.86
89
13.84
59
1.28
22
16.50
60
26.27
72
19.97
67
17.17
50
20.52
65
18.49
68
22.86
76
10.76
53
5.13
85
2.83
86
30.72
91
7.72
88
20.86
89
11.01
89
stereogantwo views7.69
47
0.88
17
7.08
45
3.49
63
3.93
66
18.98
69
3.23
32
16.52
61
19.58
61
9.93
31
18.92
54
20.50
64
9.04
51
14.07
54
6.14
29
0.26
56
0.04
37
0.21
69
0.03
45
0.63
62
0.33
46
PWC_ROBbinarytwo views8.24
52
3.13
71
12.74
63
2.43
41
4.43
71
7.51
30
1.22
20
16.63
62
19.24
60
16.08
59
28.29
67
13.99
45
10.16
55
13.63
52
14.06
63
0.42
70
0.00
1
0.05
61
0.00
1
0.59
60
0.27
44
RTSCtwo views9.15
56
3.00
70
13.57
66
3.72
66
1.76
43
11.82
50
0.46
9
16.95
63
36.83
82
15.80
58
15.53
46
12.91
39
7.46
44
20.01
73
21.76
74
0.31
61
0.13
57
0.01
39
0.08
57
0.57
58
0.41
50
AnyNet_C01two views16.12
80
10.81
87
59.36
88
4.42
72
2.49
54
30.06
82
15.15
85
17.51
64
16.51
52
17.88
62
37.69
81
24.04
69
17.54
65
29.60
83
33.29
89
0.28
60
0.38
72
0.43
78
0.42
75
2.57
82
1.98
79
FC-DCNNcopylefttwo views10.72
67
0.52
7
4.27
20
1.88
22
1.63
38
17.18
67
5.29
46
18.20
65
19.69
62
28.50
79
34.51
78
34.03
82
21.48
79
15.89
62
11.15
56
0.03
30
0.01
22
0.02
48
0.01
33
0.07
16
0.09
19
ADCStwo views13.02
71
4.93
78
28.38
82
3.17
58
2.67
55
13.61
58
10.83
74
18.70
66
33.46
79
22.59
70
24.78
62
19.59
61
18.51
69
23.40
77
32.16
88
0.10
42
0.19
65
0.37
76
0.18
63
1.26
75
1.46
78
StereoDRNettwo views5.59
31
1.75
53
6.80
43
3.12
57
4.45
72
10.61
46
4.35
41
18.80
67
9.73
25
12.22
41
6.87
1
11.44
33
4.65
26
8.09
19
8.26
39
0.02
21
0.11
54
0.00
1
0.03
45
0.20
35
0.28
45
SGM_RVCbinarytwo views10.08
62
0.60
11
3.42
10
2.30
38
0.32
5
19.41
70
6.33
52
18.95
68
14.64
43
25.14
77
24.32
61
33.34
81
18.79
70
19.86
72
12.55
60
0.25
55
0.26
69
0.22
70
0.24
68
0.34
48
0.40
49
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
SANettwo views10.64
66
1.86
57
10.91
58
1.76
18
0.71
15
14.62
60
9.23
65
19.18
69
37.14
83
19.22
65
27.96
66
25.86
73
19.11
71
13.02
48
10.63
51
0.08
40
0.06
45
0.03
54
0.02
38
0.62
61
0.81
67
PDISCO_ROBtwo views9.62
61
1.99
59
11.51
59
9.88
85
9.61
86
21.48
74
3.83
37
19.33
70
28.49
75
11.27
38
14.17
40
19.92
63
5.02
30
16.35
64
9.18
46
5.28
86
0.41
74
0.14
67
0.09
60
2.05
81
2.36
82
MDST_ROBtwo views8.37
53
0.32
1
9.03
54
4.18
70
2.42
52
26.86
80
6.14
48
19.36
71
13.52
41
27.09
78
22.75
59
9.47
25
4.74
27
15.06
59
6.34
30
0.02
21
0.02
29
0.00
1
0.00
1
0.02
4
0.13
30
SAMSARAtwo views14.63
75
2.74
67
12.38
61
12.65
87
6.74
80
36.50
86
72.93
93
19.36
71
23.77
69
16.20
60
13.04
36
29.21
77
12.78
60
16.98
65
15.21
64
0.11
43
0.26
69
0.03
54
0.14
62
0.76
65
0.77
65
SGM+DAISYtwo views15.62
78
7.26
83
19.28
75
8.94
83
10.11
87
26.25
79
10.49
72
19.36
71
14.65
44
30.64
81
33.59
77
33.00
80
22.32
80
24.96
79
16.42
67
7.90
89
6.25
90
4.51
87
3.37
86
5.86
87
7.20
86
DPSNettwo views10.14
63
1.88
58
16.82
72
1.85
20
1.73
42
24.84
78
17.20
87
19.92
74
27.41
73
12.23
43
13.62
38
16.52
51
18.35
67
14.42
57
12.50
59
0.78
75
0.54
79
0.08
63
0.25
69
1.18
73
0.59
62
DRN-Testtwo views5.87
33
0.98
24
5.89
36
2.69
48
3.65
65
12.37
51
3.35
33
20.07
75
10.20
30
11.93
40
12.31
31
11.06
32
5.31
32
7.89
13
9.05
45
0.04
33
0.05
40
0.04
59
0.04
51
0.18
33
0.25
41
SHDtwo views9.61
60
2.60
65
12.46
62
3.69
65
3.54
63
9.47
37
1.25
21
20.16
76
37.84
85
18.19
63
21.24
56
16.96
52
12.83
61
14.47
58
16.05
66
0.32
62
0.13
57
0.01
39
0.08
57
0.38
52
0.48
54
MeshStereopermissivetwo views11.52
69
1.52
48
4.55
22
1.89
23
1.46
34
19.87
72
5.11
44
20.66
77
15.91
49
32.67
84
34.51
78
39.34
87
21.15
77
18.74
70
12.10
57
0.11
43
0.06
45
0.01
39
0.00
1
0.45
57
0.22
37
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
MFMNet_retwo views13.29
73
8.60
84
18.29
73
9.75
84
7.25
83
19.65
71
14.84
83
20.71
78
30.72
77
23.03
73
28.77
68
18.85
58
26.09
84
13.55
51
9.82
48
2.44
83
1.35
84
0.34
74
0.23
67
4.78
86
6.69
85
FBW_ROBtwo views8.50
55
1.03
26
7.98
52
1.93
27
1.28
32
13.10
55
6.23
50
22.50
79
18.98
58
18.82
64
14.91
44
19.06
59
10.04
54
18.41
68
9.83
49
0.62
74
0.22
66
1.82
85
0.82
80
0.99
70
1.36
76
NVStereoNet_ROBtwo views16.04
79
6.75
82
12.90
64
6.37
81
7.42
84
12.89
53
9.74
68
22.78
80
25.12
70
30.32
80
46.19
89
34.37
83
25.38
82
21.48
74
21.38
71
5.94
87
3.10
87
6.07
88
10.09
90
4.01
84
8.54
88
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
SPS-STEREOcopylefttwo views15.04
76
6.23
81
13.21
65
11.34
86
11.65
88
23.30
76
7.15
59
24.16
81
15.65
47
31.78
83
29.19
70
31.62
79
21.32
78
24.62
78
19.50
69
7.59
88
4.19
89
3.22
86
1.48
84
6.99
88
6.54
84
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
ELAS_RVCcopylefttwo views16.54
81
2.26
62
10.09
57
5.50
79
4.46
73
28.28
81
16.72
86
25.55
82
33.54
80
40.19
86
40.30
85
36.68
85
30.03
85
29.40
82
20.61
70
0.98
81
1.21
82
0.86
81
0.70
78
1.39
77
2.16
80
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ELAScopylefttwo views16.72
82
2.14
61
9.23
55
4.92
76
4.53
75
32.66
85
15.11
84
27.40
83
28.68
76
40.27
87
44.90
88
38.33
86
30.50
86
26.44
81
21.94
76
0.88
78
1.23
83
0.67
79
0.89
81
1.49
78
2.18
81
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SGM-ForestMtwo views16.99
84
1.08
30
5.74
33
2.12
33
0.75
16
31.63
84
12.21
76
27.80
84
32.25
78
37.88
85
39.99
84
52.96
91
35.20
89
33.60
86
24.47
79
0.26
56
0.39
73
0.31
73
0.39
74
0.26
44
0.53
57
PVDtwo views15.44
77
2.93
69
14.67
68
4.21
71
3.39
62
17.43
68
4.16
38
27.84
85
48.84
89
31.02
82
43.54
87
29.76
78
30.81
87
25.97
80
21.40
72
0.23
53
0.41
74
0.04
59
0.33
71
0.41
56
1.33
75
Nwc_Nettwo views12.96
70
2.43
64
15.29
69
4.46
73
3.56
64
24.49
77
12.36
77
27.85
86
21.14
64
14.50
54
27.22
65
22.84
68
20.00
75
31.34
84
29.17
85
0.78
75
0.12
55
0.00
1
0.01
33
0.95
69
0.63
63
Abc-Nettwo views13.06
72
3.78
76
19.11
74
4.54
74
4.15
68
20.62
73
14.20
79
27.91
87
21.69
66
19.32
66
39.81
82
25.95
74
23.31
81
17.98
67
15.83
65
0.45
71
0.14
60
0.01
39
0.08
57
1.13
72
1.27
74
MANEtwo views19.47
88
1.27
41
5.07
26
4.69
75
5.55
76
30.49
83
9.94
69
34.01
88
37.27
84
44.13
89
51.57
91
52.51
90
40.41
91
33.58
85
24.81
80
0.89
79
0.86
80
1.11
83
9.72
89
0.38
52
1.06
70
MADNet+two views27.07
89
33.84
89
90.97
93
20.14
88
7.47
85
48.43
90
47.10
89
35.43
89
36.46
81
20.11
68
30.05
71
25.29
71
35.08
88
45.50
90
50.28
91
2.13
82
2.00
85
1.19
84
0.76
79
4.71
85
4.43
83
PWCKtwo views30.53
90
44.32
91
47.25
87
29.76
90
7.23
82
40.78
87
27.10
88
44.73
90
44.32
88
47.31
90
36.37
80
47.16
88
26.05
83
41.26
89
31.87
87
21.83
90
4.03
88
29.50
90
4.67
87
27.17
90
7.80
87
edge stereotwo views42.36
91
35.18
90
61.87
89
36.69
91
34.28
90
64.01
92
49.25
90
49.10
91
51.11
90
41.69
88
62.57
92
47.20
89
43.96
92
46.98
92
45.63
90
23.51
91
25.35
91
23.07
89
25.55
91
40.35
91
39.91
91
DPSimNet_ROBtwo views53.45
92
64.73
92
44.39
86
53.97
92
45.39
92
53.66
91
54.83
91
55.15
92
57.87
92
64.16
92
50.83
90
63.40
92
53.34
93
46.45
91
65.81
92
63.13
92
26.54
92
57.94
92
51.11
92
45.52
92
50.69
92
MADNet++two views82.84
93
82.38
93
73.57
90
87.72
93
82.97
93
93.14
93
69.15
92
86.42
93
82.50
93
93.46
93
86.70
93
86.28
93
80.92
94
88.34
93
88.84
93
86.83
93
84.17
93
72.64
93
68.92
93
80.47
93
81.42
93
MEDIAN_ROBtwo views98.41
94
99.70
94
99.30
95
97.09
94
97.02
94
96.89
94
95.77
95
97.66
94
97.28
94
98.79
96
98.94
94
99.18
94
98.14
95
96.89
94
96.88
94
99.96
96
99.16
94
100.00
94
99.99
94
99.69
94
99.88
94
DGTPSM_ROBtwo views99.90
96
100.00
96
99.99
96
99.99
97
100.00
95
100.00
96
100.00
96
99.97
95
100.00
95
98.35
94
100.00
95
99.84
95
100.00
96
99.98
95
99.99
95
99.99
97
100.00
95
100.00
94
100.00
95
100.00
97
100.00
98
DPSMNet_ROBtwo views99.91
97
100.00
96
99.99
96
99.99
97
100.00
95
100.00
96
100.00
96
99.98
96
100.00
95
98.35
94
100.00
95
99.84
95
100.00
96
99.98
95
99.99
95
100.00
98
100.00
95
100.00
94
100.00
95
100.00
97
100.00
98
DPSMtwo views99.95
98
100.00
96
100.00
98
99.76
95
100.00
95
100.00
96
100.00
96
100.00
97
100.00
95
100.00
97
100.00
95
100.00
97
100.00
96
100.00
97
100.00
98
99.21
94
100.00
95
100.00
94
100.00
95
99.99
95
99.95
95
AVERAGE_ROBtwo views99.62
95
99.95
95
98.81
94
100.00
99
100.00
95
98.08
95
95.47
94
100.00
97
100.00
95
100.00
97
100.00
95
100.00
97
100.00
96
100.00
97
99.99
95
100.00
98
100.00
95
100.00
94
100.00
95
100.00
97
100.00
98
DPSM_ROBtwo views99.95
98
100.00
96
100.00
98
99.76
95
100.00
95
100.00
96
100.00
96
100.00
97
100.00
95
100.00
97
100.00
95
100.00
97
100.00
96
100.00
97
100.00
98
99.21
94
100.00
95
100.00
94
100.00
95
99.99
95
99.95
95
LSM0two views100.00
100
100.00
96
100.00
98
100.00
99
100.00
95
100.00
96
100.00
96
100.00
97
100.00
95
100.00
97
100.00
95
100.00
97
100.00
96
100.00
97
100.00
98
100.00
98
100.00
95
100.00
94
100.00
95
100.00
97
99.99
97
MSMDNettwo views1.26
4