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 Infoalldeliv. area 1ldeliv. area 1sdeliv. area 2ldeliv. area 2sdeliv. area 3ldeliv. area 3select. 1lelect. 1select. 2lelect. 2select. 3lelect. 3sfacade 1sforest 1sforest 2splayg. 1lplayg. 1splayg. 2lplayg. 2splayg. 3lplayg. 3sterra. 1sterra. 2sterra. 1lterra. 1sterra. 2lterra. 2s
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
TDLMtwo views0.09
1
0.08
3
0.12
5
0.02
1
0.02
1
0.00
1
0.01
4
0.19
4
0.04
1
0.00
1
0.00
1
0.48
8
0.01
1
0.01
2
0.26
1
0.22
1
0.08
1
0.06
4
0.01
1
0.00
1
0.10
1
0.10
4
0.01
4
0.01
1
0.00
1
0.01
1
0.30
1
0.39
7
CVANet_RVCtwo views0.21
2
0.15
4
0.61
18
0.42
17
0.06
3
0.06
16
0.01
4
1.53
21
0.05
3
0.08
4
0.04
6
0.62
13
0.04
2
0.03
5
0.46
2
0.32
2
0.14
4
0.09
7
0.05
2
0.01
2
0.19
3
0.07
3
0.00
1
0.03
4
0.00
1
0.01
1
0.55
3
0.08
1
AANet_RVCtwo views0.25
3
0.05
2
0.29
8
0.06
3
0.19
6
0.00
1
0.23
25
0.03
2
0.17
8
0.35
11
0.01
5
0.07
2
0.04
2
0.01
2
1.37
5
0.48
4
0.10
2
0.06
4
0.43
5
0.17
8
0.10
1
0.25
5
0.03
10
0.05
5
0.02
3
0.06
4
1.15
5
0.89
14
GANettwo views0.29
4
0.56
18
0.39
15
0.27
9
0.11
5
0.88
32
0.14
20
0.27
7
0.06
4
0.02
2
0.28
11
1.85
27
0.04
2
0.03
5
0.62
3
0.39
3
0.37
12
0.14
10
0.24
3
0.01
2
0.20
4
0.05
2
0.06
15
0.02
2
0.06
5
0.01
1
0.48
2
0.16
3
RYNettwo views0.45
5
0.04
1
0.00
1
0.02
1
0.03
2
0.00
1
0.00
1
0.01
1
0.04
1
0.07
3
0.00
1
0.05
1
0.12
7
0.00
1
4.77
38
0.60
5
0.10
2
1.00
41
0.35
4
0.01
2
0.21
5
0.04
1
0.00
1
0.02
2
0.08
6
3.06
27
1.39
6
0.13
2
CFNettwo views0.57
6
0.46
11
0.64
19
0.43
18
0.34
15
0.04
15
0.00
1
0.58
9
0.09
5
0.66
15
0.05
7
0.64
14
0.07
5
0.05
10
2.25
7
0.65
6
0.35
8
0.06
4
0.73
8
0.34
14
0.85
11
0.73
6
0.00
1
0.08
7
0.45
18
0.16
5
3.22
15
1.51
20
HITNettwo views0.59
7
0.16
5
0.10
4
0.06
3
0.07
4
0.01
7
0.02
9
1.76
25
0.82
19
0.11
5
0.00
1
0.68
17
0.28
16
0.11
22
2.51
10
1.02
10
0.25
7
0.13
8
0.63
6
0.24
12
1.43
19
1.56
14
0.01
4
0.12
10
0.02
3
0.16
5
3.32
16
0.42
8
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
RASNettwo views0.63
8
0.18
6
0.57
17
0.41
16
0.28
9
0.10
18
0.05
15
0.94
13
0.14
7
1.07
23
0.09
9
0.48
8
0.34
19
0.02
4
1.28
4
1.47
23
0.35
8
0.21
12
0.99
9
0.02
5
1.15
14
1.53
12
0.10
17
0.34
23
0.22
12
0.30
9
3.55
20
0.73
11
DISCOtwo views0.79
9
0.95
29
0.31
10
0.06
3
0.33
13
0.51
26
0.44
28
0.23
6
0.32
13
1.75
32
0.59
17
0.38
7
0.29
17
0.04
8
2.98
13
1.36
17
0.96
28
0.29
17
3.24
28
0.53
23
0.79
10
1.45
10
0.10
17
0.07
6
0.12
9
0.29
8
2.76
11
0.31
6
NVstereo2Dtwo views0.83
10
0.43
10
0.07
3
0.23
8
0.36
16
0.41
25
0.08
16
1.61
23
0.59
16
0.84
17
1.11
24
0.36
5
0.21
14
0.07
15
5.88
47
2.16
36
0.62
18
0.45
22
1.50
13
0.19
9
1.20
15
0.94
9
0.02
6
0.32
19
1.30
26
0.24
7
0.95
4
0.24
5
CFNet_RVCtwo views0.86
11
0.59
19
0.74
23
0.33
11
0.32
12
0.11
19
0.02
9
0.11
3
0.17
8
0.87
20
0.74
20
0.57
12
0.16
10
0.08
18
3.11
15
1.19
16
0.50
14
0.34
19
1.28
11
0.15
6
1.68
22
1.48
11
0.02
6
0.25
12
0.86
24
1.89
22
3.64
22
2.13
24
DLCB_ROBtwo views0.92
12
0.72
23
0.64
19
0.43
18
0.42
17
0.14
21
0.18
23
0.43
8
0.28
11
0.85
18
0.53
14
1.12
23
0.72
35
0.07
15
2.16
6
1.12
13
0.88
26
0.26
14
0.69
7
0.20
10
1.35
18
1.53
12
0.49
43
0.28
15
4.11
47
2.41
24
2.04
9
0.78
12
HSMtwo views0.97
13
0.30
7
0.87
26
0.54
23
0.46
20
0.00
1
0.01
4
2.93
43
0.30
12
0.13
6
0.56
16
0.65
15
0.13
8
0.05
10
2.97
12
1.08
12
0.66
21
0.17
11
3.26
30
0.35
15
1.78
23
2.03
19
0.04
11
0.33
22
0.38
15
0.81
17
4.16
34
1.10
16
NLCA_NET_v2_RVCtwo views0.99
14
0.82
26
0.34
13
0.18
7
0.43
18
0.03
13
1.10
38
1.20
17
2.50
46
2.06
38
0.11
10
0.69
18
0.07
5
0.54
39
3.65
25
0.91
7
0.19
6
0.22
13
1.65
17
0.15
6
0.73
9
0.92
8
0.08
16
0.18
11
0.11
8
3.29
29
2.58
10
2.09
23
Zhibo Rao, Mingyi He, Yuchao Dai, Zhidong Zhu, Bo Li, and Renjie He.: NLCA-Net: A non-local context attention network for stereo matching.
PDISCO_ROBtwo views1.03
15
0.55
17
0.06
2
0.67
24
0.28
9
0.00
1
0.21
24
0.71
11
0.49
15
1.28
25
0.99
23
0.25
4
0.32
18
0.19
27
3.02
14
2.03
33
0.69
22
1.42
50
2.26
18
0.58
25
2.32
27
3.22
34
0.14
20
0.27
14
1.36
27
0.37
10
3.84
28
0.23
4
iResNet_ROBtwo views1.07
16
0.65
22
0.30
9
0.45
20
0.30
11
0.01
7
0.02
9
3.10
44
0.12
6
1.34
26
0.00
1
0.66
16
0.35
20
0.04
8
4.37
31
1.65
28
0.54
15
0.49
24
5.83
41
0.29
13
1.30
17
0.87
7
0.20
26
0.26
13
0.16
10
0.84
18
3.74
26
1.12
17
AF-Nettwo views1.11
17
0.51
13
0.87
26
0.34
12
0.78
24
0.89
33
1.87
43
2.40
33
1.15
24
0.48
13
0.68
18
3.66
35
0.16
10
0.27
31
2.44
8
0.98
8
0.35
8
0.01
1
1.60
14
0.40
16
0.50
6
2.58
25
0.05
13
0.08
7
0.66
21
0.45
12
3.47
18
2.34
25
Nwc_Nettwo views1.11
17
0.51
13
0.87
26
0.34
12
0.78
24
0.89
33
1.87
43
2.40
33
1.15
24
0.48
13
0.68
18
3.66
35
0.16
10
0.27
31
2.44
8
0.98
8
0.35
8
0.01
1
1.60
14
0.40
16
0.50
6
2.58
25
0.05
13
0.08
7
0.66
21
0.45
12
3.47
18
2.34
25
iResNetv2_ROBtwo views1.26
19
0.75
24
0.32
11
1.29
40
0.33
13
0.02
12
0.02
9
3.19
45
0.20
10
2.00
37
0.29
12
2.24
30
0.41
24
0.06
13
3.64
24
1.45
22
0.75
23
0.13
8
5.77
39
0.43
20
2.86
33
1.71
16
0.16
23
0.37
25
0.38
15
0.47
14
4.08
32
0.72
10
RPtwo views1.29
20
1.06
31
2.32
56
0.84
28
1.16
28
0.39
24
1.53
40
1.89
29
1.68
35
0.29
8
2.21
33
2.25
31
0.13
8
0.40
36
3.31
19
1.07
11
0.61
16
0.05
3
1.61
16
0.40
16
0.65
8
3.05
31
0.11
19
0.29
16
0.29
13
0.92
19
4.24
35
2.04
22
BEATNet_4xtwo views1.40
21
0.54
16
0.27
7
0.13
6
0.27
8
0.01
7
0.10
17
1.80
26
1.74
37
0.41
12
0.53
14
0.82
20
0.59
30
0.54
39
4.30
28
2.92
44
1.87
42
1.84
58
1.39
12
0.71
28
4.68
47
6.21
62
0.02
6
0.81
36
0.21
11
0.40
11
3.69
23
0.96
15
DN-CSS_ROBtwo views1.45
22
0.52
15
1.11
33
0.81
27
0.69
22
0.00
1
0.03
14
2.86
42
0.35
14
1.50
30
3.76
47
0.73
19
0.36
22
0.07
15
2.89
11
1.44
20
1.08
31
0.79
30
8.61
50
0.23
11
3.25
38
1.97
18
0.02
6
0.66
33
0.10
7
0.52
15
3.12
14
1.56
21
DeepPruner_ROBtwo views1.57
23
0.37
9
0.71
21
0.50
22
0.55
21
0.37
23
1.60
41
1.74
24
0.84
20
1.37
28
2.40
35
0.49
10
0.51
28
0.10
21
5.38
43
2.16
36
1.10
32
1.08
43
2.28
19
0.40
16
4.83
51
3.42
35
0.66
50
0.34
23
0.57
20
5.04
43
3.07
12
0.53
9
PA-Nettwo views1.61
24
0.81
25
1.43
43
0.27
9
1.99
47
0.12
20
0.34
26
0.19
4
1.90
39
0.85
18
0.41
13
0.24
3
0.20
13
0.78
53
4.61
36
1.44
20
0.17
5
0.26
14
1.08
10
0.53
23
1.44
20
3.18
33
0.65
48
0.44
28
0.99
25
7.71
56
1.78
8
9.78
47
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
NCC-stereotwo views1.86
25
1.63
37
1.33
39
0.95
34
1.22
32
0.80
29
2.29
49
2.62
39
2.93
49
0.34
9
2.87
42
2.16
28
1.17
39
0.21
28
3.63
21
1.13
14
0.63
19
0.96
39
2.79
24
0.75
29
0.98
12
2.96
29
0.36
36
0.32
19
2.31
35
4.50
39
3.69
23
4.78
33
Abc-Nettwo views1.86
25
1.63
37
1.33
39
0.95
34
1.22
32
0.80
29
2.29
49
2.62
39
2.93
49
0.34
9
2.87
42
2.16
28
1.17
39
0.21
28
3.63
21
1.13
14
0.63
19
0.96
39
2.79
24
0.75
29
0.98
12
2.96
29
0.36
36
0.32
19
2.31
35
4.50
39
3.69
23
4.78
33
stereogantwo views1.89
27
2.00
45
1.17
35
1.47
42
2.05
50
1.72
41
0.87
34
1.84
28
1.11
22
1.36
27
1.51
27
5.81
44
0.48
27
0.32
35
3.87
26
1.69
29
0.80
24
0.39
21
3.96
33
0.78
31
1.23
16
4.02
42
0.26
29
0.51
31
0.30
14
1.30
21
6.30
48
3.84
31
NCCL2two views2.12
28
2.27
48
1.29
36
1.28
39
1.85
44
0.18
22
0.77
33
1.24
18
1.25
27
1.62
31
2.71
39
1.06
21
0.51
28
0.31
33
4.40
32
2.33
40
1.57
38
0.62
27
2.35
20
0.68
27
2.68
32
4.29
46
1.52
62
1.06
38
4.65
51
4.17
35
4.04
31
6.58
39
R-Stereo Traintwo views2.16
29
1.08
32
1.57
47
0.84
28
0.19
6
0.01
7
0.01
4
2.05
32
1.01
21
1.37
28
2.45
36
1.50
25
0.46
25
0.21
28
3.23
17
4.53
59
1.62
39
1.05
42
7.07
44
1.09
39
4.86
52
5.93
60
0.04
11
0.31
17
2.48
38
4.63
41
3.10
13
5.74
37
GANetREF_RVCpermissivetwo views2.43
30
1.69
40
0.81
24
2.04
48
1.16
28
1.02
36
0.10
17
4.72
60
1.44
30
3.62
57
0.06
8
10.03
54
0.69
33
0.09
19
3.63
21
1.36
17
1.06
30
0.94
37
6.44
43
0.46
21
2.32
27
1.74
17
0.17
24
1.35
45
3.41
45
3.75
33
6.55
49
4.82
35
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
RGCtwo views2.44
31
1.98
44
4.87
68
1.54
43
1.65
38
1.74
42
1.05
36
2.02
31
1.43
29
0.88
21
4.42
50
2.52
33
2.42
46
0.58
44
4.33
29
1.63
27
0.91
27
0.48
23
4.28
34
1.08
38
1.94
24
4.01
41
0.15
21
0.72
35
2.19
33
5.22
45
4.90
41
6.96
40
MLCVtwo views2.45
32
0.61
20
0.81
24
2.48
51
1.21
31
0.95
35
0.50
29
3.57
49
0.61
17
3.75
60
4.26
48
5.13
41
3.34
49
0.06
13
4.63
37
1.78
31
1.32
35
0.86
34
10.05
52
1.09
39
2.30
26
3.92
39
0.29
30
2.16
51
1.87
30
3.06
27
4.01
30
1.40
18
R-Stereotwo views2.54
33
0.34
8
1.32
37
0.35
14
0.45
19
0.01
7
0.17
21
3.20
46
1.38
28
1.80
34
0.85
21
1.70
26
0.47
26
0.17
25
3.19
16
5.06
65
1.40
37
0.59
26
10.54
53
1.73
48
6.89
64
5.24
52
0.20
26
0.41
26
2.41
37
4.21
36
5.91
45
8.54
46
RTSCtwo views2.56
34
0.89
28
1.50
44
0.47
21
0.76
23
0.87
31
2.29
49
1.56
22
1.91
40
2.12
39
2.86
41
1.23
24
0.97
38
1.03
62
7.27
60
3.36
52
4.68
63
1.74
55
7.73
46
1.11
41
3.95
42
5.39
55
0.21
28
1.11
41
0.40
17
0.95
20
4.56
39
8.26
45
DeepPrunerFtwo views2.66
35
1.67
39
0.33
12
3.58
58
9.30
76
1.76
43
3.48
58
0.68
10
3.29
52
0.71
16
1.59
29
0.37
6
0.27
15
0.13
23
5.94
48
3.42
55
1.03
29
0.54
25
4.47
35
0.65
26
2.02
25
2.06
20
0.56
46
0.31
17
15.21
74
3.30
30
3.74
26
1.47
19
NOSS_ROBtwo views2.78
36
3.78
61
0.39
15
2.02
47
1.58
37
2.36
48
0.00
1
2.66
41
1.54
32
3.63
58
1.79
32
2.55
34
3.51
50
0.83
56
5.82
46
5.57
67
2.69
48
1.83
57
2.63
23
0.91
35
4.62
45
4.13
43
0.42
39
2.58
55
2.29
34
4.46
38
7.67
56
2.88
27
SGM-Foresttwo views2.81
37
2.03
46
0.36
14
6.32
69
1.79
43
2.66
49
0.02
9
1.42
19
0.77
18
2.52
40
0.96
22
6.57
46
5.74
53
0.31
33
4.36
30
2.94
46
0.61
16
0.33
18
3.08
26
0.47
22
2.46
30
2.73
27
0.84
55
0.45
29
3.66
46
3.74
32
11.30
62
7.44
42
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
PWCDC_ROBbinarytwo views2.82
38
0.49
12
0.17
6
0.38
15
0.79
26
0.03
13
0.10
17
1.16
16
8.93
67
0.14
7
2.49
37
0.55
11
31.02
80
0.57
41
5.58
45
3.32
51
1.73
40
0.91
36
2.53
22
0.88
33
2.92
34
7.18
66
0.15
21
0.53
32
0.55
19
0.69
16
1.43
7
0.88
13
HSM-Net_RVCpermissivetwo views2.86
39
1.24
34
1.50
44
2.55
53
1.72
39
0.57
27
0.17
21
3.55
48
1.14
23
3.50
56
6.32
61
7.94
47
2.93
47
0.05
10
3.38
20
2.12
34
1.28
34
0.26
14
3.24
28
1.24
43
3.66
40
3.76
38
0.61
47
2.64
56
4.43
49
6.27
51
5.76
44
5.47
36
Gengshan Yang, Joshua Manela, Michael Happold, and Deva Ramanan: Hierarchical Deep Stereo Matching on High-resolution Images. CVPR 2019
ADCReftwo views2.96
40
2.78
55
1.02
31
0.70
25
2.21
51
1.96
45
2.42
52
0.96
14
6.10
60
1.94
36
1.15
25
1.08
22
0.74
36
1.32
67
6.01
49
3.38
53
1.34
36
1.15
45
8.41
49
2.09
53
2.59
31
3.60
37
1.88
64
0.71
34
2.81
41
8.00
58
3.42
17
10.24
48
NaN_ROBtwo views3.01
41
1.16
33
1.13
34
0.86
31
3.03
56
0.59
28
0.75
31
3.25
47
2.58
47
0.98
22
1.32
26
21.10
65
0.59
30
1.15
65
4.15
27
1.58
26
0.80
24
0.70
28
2.38
21
1.30
44
1.52
21
2.15
21
0.46
42
1.07
40
2.50
39
6.25
50
4.54
38
13.27
53
XPNet_ROBtwo views3.02
42
2.54
51
1.33
39
0.73
26
1.45
36
1.04
37
0.61
30
0.83
12
1.45
31
1.16
24
3.44
44
26.85
70
0.35
20
0.40
36
3.24
18
2.22
39
3.28
53
0.79
30
3.20
27
1.19
42
2.94
36
5.51
56
0.76
52
0.46
30
0.72
23
2.37
23
9.22
58
3.37
30
ccstwo views3.29
43
0.63
21
1.94
51
0.99
36
2.02
48
0.06
16
0.01
4
3.87
52
1.79
38
4.48
72
9.13
72
4.94
39
4.74
52
0.87
57
6.01
49
5.69
68
1.83
41
1.39
48
5.37
37
1.03
37
5.86
59
7.35
68
0.32
33
1.56
46
3.32
44
6.28
52
4.08
32
3.36
29
ADCP+two views3.47
44
2.05
47
2.64
59
1.31
41
1.93
45
1.19
38
3.99
59
1.07
15
9.83
69
1.78
33
5.75
57
8.79
51
0.37
23
0.90
58
5.04
42
2.20
38
0.44
13
0.90
35
5.82
40
0.88
33
2.38
29
3.43
36
0.33
34
0.41
26
3.03
43
12.72
68
3.59
21
10.90
49
FC-DCNNcopylefttwo views3.82
45
4.38
65
5.67
75
9.61
73
4.14
60
8.01
70
3.26
56
4.21
53
1.69
36
2.95
48
1.54
28
5.87
45
2.09
45
0.13
23
4.40
32
2.15
35
2.54
47
1.11
44
3.38
31
0.82
32
3.11
37
2.20
22
0.65
48
3.29
58
6.33
61
4.91
42
6.98
52
7.80
44
ADCLtwo views3.91
46
3.61
59
2.28
54
1.76
45
1.76
42
3.09
52
4.16
60
1.98
30
13.71
76
2.81
46
5.14
53
2.47
32
0.65
32
1.10
64
5.38
43
2.55
42
2.02
43
2.20
61
6.41
42
2.08
52
4.87
53
4.20
44
1.40
61
1.06
38
6.69
62
6.60
53
4.33
37
11.16
51
LALA_ROBtwo views4.04
47
4.73
67
1.42
42
1.16
38
1.93
45
2.17
47
1.04
35
1.80
26
2.48
45
2.74
43
5.87
58
33.34
72
0.69
33
0.78
53
6.02
51
2.33
40
3.21
52
1.82
56
3.66
32
1.88
49
3.40
39
4.99
49
0.34
35
1.03
37
1.91
31
3.61
31
8.50
57
6.20
38
iResNettwo views4.14
48
0.84
27
0.71
21
2.34
50
2.59
52
4.58
59
1.07
37
4.67
58
1.21
26
3.72
59
3.61
45
12.28
56
16.06
64
0.03
5
6.64
57
3.14
49
2.35
46
1.89
59
16.43
60
2.53
57
3.76
41
2.90
28
0.17
24
3.88
63
1.66
29
4.23
37
5.19
42
3.23
28
SHDtwo views4.34
49
1.02
30
2.50
57
1.71
44
1.33
34
1.46
39
2.01
45
4.90
61
4.30
58
2.63
41
4.99
52
4.39
38
9.41
57
0.58
44
8.20
71
4.81
61
4.43
62
2.44
65
13.01
58
2.95
62
5.61
58
6.88
65
0.30
31
3.26
57
4.11
47
5.16
44
3.88
29
10.98
50
CSANtwo views4.46
50
2.36
50
1.57
47
3.12
55
3.32
57
1.64
40
0.76
32
4.59
57
4.37
59
3.80
62
2.84
40
17.58
62
10.47
59
0.96
60
4.80
40
1.75
30
2.06
44
1.16
46
5.69
38
2.34
55
2.93
35
2.40
24
0.43
40
2.56
54
4.64
50
11.45
66
7.53
55
13.44
54
ETE_ROBtwo views4.62
51
4.04
63
1.52
46
1.14
37
2.02
48
2.70
50
1.28
39
1.50
20
4.25
57
1.89
35
5.88
59
37.95
77
0.84
37
1.42
70
4.94
41
2.66
43
3.81
58
0.95
38
4.84
36
1.63
47
4.01
43
5.32
54
0.67
51
1.14
42
2.03
32
3.84
34
15.25
69
7.23
41
XQCtwo views4.76
52
3.14
57
1.63
49
3.31
56
3.69
59
3.46
55
2.94
55
6.02
67
3.05
51
2.64
42
5.51
55
5.67
43
7.46
54
1.41
69
8.02
65
4.65
60
3.59
55
1.49
52
10.75
54
2.91
58
4.96
55
6.67
64
0.41
38
1.91
47
7.13
66
10.47
65
11.17
61
4.51
32
MSMDNettwo views4.98
53
2.59
53
3.57
62
2.54
52
0.90
27
3.99
56
1.63
42
4.52
56
1.56
33
4.54
73
18.92
89
20.82
64
12.09
60
0.60
48
6.57
56
5.84
69
2.28
45
1.52
53
7.08
45
1.94
51
6.09
60
6.12
61
0.31
32
2.29
52
1.47
28
2.83
26
4.25
36
7.49
43
CBMVpermissivetwo views5.52
54
3.65
60
1.32
37
3.46
57
2.86
55
2.04
46
0.41
27
3.64
50
2.63
48
4.09
66
2.36
34
14.13
59
3.70
51
0.57
41
7.16
59
6.19
72
4.26
60
2.36
64
8.76
51
1.44
45
6.43
61
4.53
48
0.54
45
5.62
72
4.85
52
6.08
48
19.41
73
26.62
75
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
RTStwo views5.53
55
1.46
35
0.94
29
0.90
32
1.73
40
4.08
57
4.63
64
2.53
36
3.41
53
2.79
44
1.70
30
8.18
48
1.27
41
0.72
49
7.83
62
4.87
62
7.37
75
2.34
62
18.33
61
2.92
59
4.78
48
5.88
58
1.04
56
1.31
43
6.74
63
22.88
86
6.66
50
22.00
65
RTSAtwo views5.53
55
1.46
35
0.94
29
0.90
32
1.73
40
4.08
57
4.63
64
2.53
36
3.41
53
2.79
44
1.70
30
8.18
48
1.27
41
0.72
49
7.83
62
4.87
62
7.37
75
2.34
62
18.33
61
2.92
59
4.78
48
5.88
58
1.04
56
1.31
43
6.74
63
22.88
86
6.66
50
22.00
65
AnyNet_C32two views5.68
57
1.96
42
2.91
60
1.78
46
2.71
54
3.17
53
5.76
69
4.31
54
12.56
71
3.07
50
7.17
63
4.22
37
1.70
43
2.19
72
8.22
72
5.98
70
5.69
68
4.43
75
12.45
57
2.93
61
9.72
74
8.57
76
1.84
63
3.46
60
6.24
60
9.82
63
6.26
47
14.21
55
CBMV_ROBtwo views6.09
58
2.65
54
2.29
55
8.35
72
5.74
66
5.08
60
2.76
54
5.03
63
2.35
43
3.78
61
4.70
51
21.22
66
13.56
62
0.09
19
6.04
52
3.70
57
2.84
49
1.39
48
11.35
55
8.80
79
4.92
54
3.11
32
0.43
40
2.10
50
7.06
65
10.12
64
12.96
64
11.93
52
ADCPNettwo views6.67
59
1.97
43
3.86
64
2.75
54
6.19
70
3.07
51
7.60
75
5.21
64
8.03
65
3.82
63
7.92
66
5.17
42
13.78
63
2.98
76
8.10
67
6.12
71
4.35
61
2.04
60
18.50
63
4.74
69
7.04
65
7.43
69
2.98
71
1.91
47
8.96
69
14.04
73
7.02
54
14.47
56
AMNettwo views6.77
60
3.87
62
2.62
58
4.00
62
5.86
67
6.75
66
3.31
57
2.51
35
2.47
44
3.44
54
6.68
62
9.34
52
2.05
44
7.44
87
4.79
39
9.81
83
13.48
84
7.65
84
7.95
48
7.83
74
11.47
78
7.62
71
2.74
69
4.18
64
5.52
55
8.25
60
13.24
65
17.97
61
ADCMidtwo views6.83
61
4.46
66
4.95
69
2.16
49
4.31
63
5.71
64
7.29
73
8.83
68
12.78
73
3.11
51
8.10
68
9.83
53
9.65
58
2.84
73
8.51
74
5.42
66
4.68
63
2.83
67
18.87
65
3.22
63
7.09
66
7.70
72
5.37
78
2.33
53
5.12
54
8.71
61
5.21
43
15.24
57
PWC_ROBbinarytwo views7.15
62
1.85
41
5.57
73
0.85
30
1.17
30
1.95
44
4.77
66
2.56
38
17.27
86
3.83
64
7.66
65
4.97
40
39.20
88
0.81
55
8.35
73
4.96
64
6.51
71
4.27
72
15.99
59
4.39
66
7.48
68
7.59
70
0.53
44
4.79
69
2.74
40
2.63
25
7.01
53
23.26
68
FBW_ROBtwo views7.98
63
9.24
77
5.57
73
20.21
80
8.25
74
12.08
75
8.09
76
4.92
62
1.58
34
3.30
53
5.15
54
13.88
58
3.03
48
7.45
88
6.23
53
1.49
24
2.89
50
0.79
30
33.18
82
1.93
50
5.24
56
4.40
47
9.43
84
8.29
80
4.88
53
7.69
55
4.88
40
21.40
63
pmcnntwo views8.05
64
2.96
56
2.25
53
6.23
67
5.03
64
6.98
67
4.33
62
4.70
59
2.02
42
4.19
67
9.20
74
19.11
63
20.39
70
0.73
51
7.05
58
6.73
73
6.29
70
4.51
76
22.54
72
4.66
68
9.62
73
5.08
51
4.46
77
4.23
65
5.99
56
5.22
45
17.26
71
25.55
74
ADCStwo views9.93
65
5.64
70
4.82
67
3.95
61
6.05
69
6.17
65
5.70
68
9.47
70
15.68
77
4.19
67
7.98
67
12.91
57
17.41
65
3.94
83
10.52
82
11.68
87
10.14
77
7.14
83
24.46
76
8.67
78
13.95
84
13.00
87
2.94
70
6.15
75
9.75
71
17.65
78
10.70
60
17.41
59
SGM_RVCbinarytwo views10.08
66
10.50
79
3.60
63
16.68
76
8.80
75
14.87
76
7.11
72
9.19
69
6.25
61
4.06
65
7.33
64
24.89
67
18.29
67
0.57
41
4.57
35
1.98
32
3.30
54
1.17
47
20.75
68
2.20
54
4.80
50
2.23
23
2.29
66
3.65
61
17.68
79
13.31
70
33.30
83
28.67
77
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
MDST_ROBtwo views10.13
67
16.70
85
3.04
61
29.06
84
9.34
77
20.66
82
5.02
67
5.90
66
1.98
41
3.16
52
2.70
38
11.22
55
28.15
76
0.18
26
6.30
54
1.37
19
1.21
33
0.34
19
7.74
47
1.01
36
4.67
46
3.95
40
1.92
65
1.94
49
26.55
83
19.78
80
37.02
86
22.50
67
SANettwo views10.32
68
4.15
64
6.27
76
5.51
65
5.49
65
8.58
72
2.27
48
15.41
81
13.03
75
2.83
47
9.09
71
60.35
91
25.50
73
2.93
75
8.06
66
3.02
48
7.05
72
3.86
71
11.43
56
3.25
64
5.27
57
5.02
50
2.60
68
7.27
78
11.08
72
12.57
67
13.36
66
23.31
69
ccs_robtwo views10.35
69
5.69
71
2.14
52
4.64
64
2.67
53
5.53
63
6.85
71
4.40
55
4.18
56
5.98
78
10.34
76
26.00
68
37.16
86
3.25
79
19.47
92
10.00
84
11.64
82
3.40
69
48.57
87
13.46
83
7.76
69
10.45
80
0.80
53
4.63
67
2.91
42
6.10
49
6.08
46
15.43
58
STTStereo_v2two views10.87
70
4.96
68
9.71
79
3.77
59
4.24
61
7.37
68
2.01
45
10.08
71
15.80
79
4.95
75
15.59
82
48.52
86
9.05
55
0.58
44
8.11
69
8.67
78
5.49
65
4.42
73
23.80
73
15.66
86
8.01
70
8.23
74
3.06
73
5.14
70
6.15
58
13.92
71
22.56
75
23.75
70
G-Nettwo views10.87
70
4.96
68
9.71
79
3.77
59
4.24
61
7.37
68
2.01
45
10.08
71
15.80
79
4.95
75
15.59
82
48.52
86
9.05
55
0.58
44
8.11
69
8.67
78
5.49
65
4.42
73
23.80
73
15.66
86
8.01
70
8.23
74
3.06
73
5.14
70
6.15
58
13.92
71
22.56
75
23.75
70
edge stereotwo views11.35
72
6.83
75
10.52
81
14.06
74
12.00
81
16.55
78
12.81
80
12.42
76
7.44
63
6.06
79
15.12
81
38.81
79
17.75
66
1.25
66
7.59
61
7.29
75
5.73
69
3.11
68
25.18
77
4.53
67
6.71
62
8.01
73
4.16
75
3.71
62
8.94
68
7.65
54
13.41
67
28.72
78
SGM-ForestMtwo views11.60
73
15.35
83
4.54
65
19.02
78
8.08
73
18.74
81
10.83
79
12.03
73
7.24
62
3.01
49
6.09
60
36.06
76
23.95
71
0.75
52
4.47
34
1.51
25
3.60
56
0.75
29
20.06
67
1.60
46
4.23
44
1.65
15
1.36
60
3.34
59
25.09
82
20.48
82
34.27
85
25.03
73
AnyNet_C01two views12.12
74
6.13
74
5.39
71
6.24
68
6.03
68
5.20
61
9.20
77
12.30
75
16.87
83
5.85
77
9.03
70
26.16
69
18.91
68
4.53
85
11.14
85
12.98
88
16.27
85
15.82
90
34.24
83
10.36
81
16.44
85
13.34
88
7.06
79
6.93
77
11.61
73
9.46
62
9.77
59
19.91
62
PVDtwo views12.14
75
5.88
72
8.51
78
7.70
71
6.65
71
10.05
73
6.16
70
21.11
89
19.74
89
3.44
54
8.69
69
16.07
60
27.50
75
2.86
74
8.93
76
7.52
77
11.20
80
6.11
81
24.24
75
6.83
73
11.08
75
11.21
82
1.11
58
5.71
73
15.30
75
21.90
84
17.71
72
34.49
83
MeshStereopermissivetwo views13.21
76
10.28
78
5.53
72
24.29
83
14.71
82
17.26
79
7.30
74
12.03
73
7.47
64
4.66
74
9.13
72
38.56
78
29.80
77
0.45
38
9.43
78
7.37
76
3.76
57
2.44
65
18.85
64
2.41
56
12.93
82
7.18
66
1.21
59
5.79
74
22.04
81
16.38
76
33.15
82
32.34
81
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
DPSNettwo views14.25
77
3.23
58
1.83
50
7.63
70
7.22
72
10.44
74
4.19
61
5.68
65
8.64
66
4.20
69
3.75
46
72.37
93
26.68
74
4.43
84
10.42
80
10.25
86
10.56
78
8.02
85
58.50
91
8.44
77
11.36
77
8.93
77
16.95
89
10.34
85
18.62
80
15.80
75
14.33
68
21.91
64
MFMNet_retwo views16.26
78
13.95
81
7.70
77
24.27
82
9.34
77
33.99
89
17.43
82
12.42
76
19.62
88
10.23
86
13.58
80
39.02
80
13.47
61
1.40
68
10.73
83
10.12
85
29.43
91
12.22
87
31.31
79
10.07
80
17.19
86
9.15
78
8.44
82
9.02
82
8.97
70
5.68
47
24.08
77
36.04
85
STStereotwo views19.12
79
25.20
89
16.96
91
40.13
89
26.93
89
32.21
87
21.77
88
19.35
87
16.49
82
7.75
83
17.60
86
46.83
84
36.96
85
1.06
63
8.10
67
2.99
47
3.81
58
1.57
54
31.85
80
5.18
71
6.77
63
5.31
53
3.04
72
9.71
84
29.57
84
23.01
88
40.22
87
35.81
84
ELAS_RVCcopylefttwo views19.26
80
19.85
87
14.31
86
33.92
86
22.40
85
33.39
88
19.24
86
15.07
79
10.89
70
8.20
84
17.30
85
54.84
90
32.79
82
3.23
78
8.72
75
3.46
56
11.25
81
6.19
82
19.97
66
7.86
75
12.63
81
12.02
85
9.45
85
10.70
86
30.22
86
22.17
85
41.23
89
38.75
88
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SAMSARAtwo views19.29
81
2.55
52
5.07
70
4.09
63
3.44
58
3.26
54
4.38
63
15.32
80
9.58
68
4.44
70
5.63
56
8.73
50
24.82
72
3.30
81
10.45
81
6.81
74
5.49
65
3.46
70
21.26
69
4.77
70
17.96
88
24.74
93
0.82
54
4.77
68
80.12
96
91.25
96
77.91
95
76.45
95
MADNet+two views19.30
82
7.92
76
15.32
88
17.51
77
11.42
80
8.09
71
18.41
84
28.03
92
17.21
85
12.46
87
12.08
77
33.63
73
30.37
78
9.62
89
18.64
91
14.89
90
25.78
89
10.10
86
40.05
85
16.50
88
17.34
87
12.76
86
9.48
86
18.60
91
35.02
92
20.22
81
21.19
74
38.41
86
ELAScopylefttwo views19.48
83
22.90
88
14.50
87
35.26
88
24.25
88
31.48
86
19.80
87
16.72
83
15.91
81
7.71
82
15.62
84
48.88
88
30.47
79
3.20
77
9.19
77
3.39
54
10.97
79
5.79
79
21.49
70
6.53
72
11.69
79
11.90
84
9.30
83
9.63
83
32.86
87
24.17
89
41.19
88
41.31
89
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
MANEtwo views19.83
84
17.14
86
15.43
89
34.46
87
23.74
87
25.27
84
17.85
83
19.50
88
17.04
84
6.13
80
12.40
78
41.03
81
35.23
84
3.36
82
7.87
64
8.98
80
7.09
73
29.77
93
21.81
71
3.71
65
9.51
72
5.54
57
2.50
67
8.83
81
34.12
89
27.95
91
51.25
90
47.85
91
SPS-STEREOcopylefttwo views20.13
85
14.92
82
11.88
82
19.84
79
20.39
83
18.24
80
9.27
78
16.06
82
19.74
89
19.91
91
20.06
91
33.66
74
37.68
87
7.23
86
13.17
88
16.66
92
26.75
90
16.01
91
32.30
81
23.65
92
30.25
93
22.84
92
13.67
87
13.47
89
16.56
77
15.26
74
29.55
80
24.45
72
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
SGM+DAISYtwo views20.91
86
16.10
84
12.43
83
22.81
81
21.80
84
24.45
83
18.63
85
17.89
84
12.77
72
18.42
90
19.55
90
30.39
71
40.48
89
11.67
90
12.56
87
15.10
91
25.73
88
12.69
88
34.93
84
21.56
91
26.12
92
22.15
91
21.16
90
12.73
88
17.44
78
13.30
69
33.35
84
28.26
76
LE_ROBtwo views22.46
87
10.87
80
12.67
84
32.44
85
37.93
90
31.36
85
33.43
90
23.22
90
21.14
92
6.73
81
17.81
87
52.15
89
50.94
94
1.56
71
6.43
55
2.92
44
3.14
51
1.46
51
30.14
78
8.43
76
7.23
67
4.28
45
7.08
80
25.69
92
34.50
91
36.15
92
53.30
91
53.30
92
LSM0two views22.48
88
6.09
73
4.65
66
14.89
75
10.39
79
16.43
77
13.29
81
23.38
91
15.69
78
14.75
89
18.69
88
35.79
75
33.70
83
0.93
59
73.52
95
85.94
95
13.36
83
5.62
78
55.67
90
16.83
89
21.95
90
18.82
90
4.28
76
8.01
79
16.44
76
16.43
77
28.65
78
32.66
82
BEATNet-Init1two views23.62
89
27.86
90
13.22
85
44.78
90
23.54
86
38.58
90
23.96
89
18.57
85
18.15
87
9.17
85
13.26
79
47.63
85
48.00
93
3.29
80
12.31
86
9.22
81
7.35
74
5.48
77
48.70
88
19.34
90
12.33
80
11.24
83
7.11
81
11.28
87
33.76
88
21.63
83
60.52
93
47.50
90
DispFullNettwo views26.27
90
39.81
91
17.89
92
52.75
91
38.74
91
45.03
91
55.21
92
13.74
78
12.89
74
13.75
88
9.88
75
43.75
83
32.46
81
86.51
95
11.01
84
13.68
89
29.48
92
16.23
92
51.78
89
14.64
84
13.87
83
10.46
81
14.59
88
6.22
76
8.62
67
8.03
59
16.08
70
32.10
80
PSMNet_ROBtwo views31.36
91
48.75
92
16.32
90
64.48
95
46.97
93
60.77
94
61.30
93
19.02
86
20.70
91
27.57
93
24.90
92
41.43
82
42.18
90
83.15
94
13.95
89
4.06
58
17.90
86
5.94
80
42.51
86
15.37
85
11.21
76
9.54
79
57.95
94
14.27
90
29.64
85
19.74
79
29.38
79
17.81
60
PWCKtwo views38.25
92
63.47
95
23.98
93
73.60
96
41.23
92
60.98
95
51.31
91
54.00
94
35.86
93
26.51
92
39.23
93
62.00
92
46.09
91
26.43
91
14.46
90
9.53
82
30.80
93
12.71
89
83.96
94
28.46
93
23.42
91
13.60
89
47.98
92
30.51
93
34.48
90
27.91
90
31.53
81
38.61
87
DPSimNet_ROBtwo views58.77
93
55.98
93
59.21
95
62.88
94
73.92
94
60.20
93
64.84
95
28.34
93
91.63
96
80.59
95
52.65
94
74.49
94
47.24
92
90.64
96
55.91
94
45.99
94
54.62
94
42.30
94
68.03
92
59.25
94
51.18
94
59.54
95
44.54
91
45.19
94
44.85
93
61.56
93
56.15
92
55.05
93
CC-Net-ROBtwo views70.08
94
57.59
94
33.20
94
55.61
92
76.68
95
84.28
96
62.79
94
83.71
96
83.30
95
76.21
94
74.67
95
74.97
95
61.84
95
62.54
93
41.03
93
41.52
93
86.04
95
88.34
96
89.47
95
90.97
95
75.66
95
54.62
94
53.92
93
50.83
95
74.81
95
79.04
95
83.52
96
94.87
97
MADNet++two views76.54
95
72.15
96
83.11
96
61.53
93
81.21
96
58.42
92
69.47
96
68.24
95
78.58
94
85.63
96
92.37
96
80.60
96
85.04
96
36.71
92
81.19
96
89.82
96
88.15
96
69.64
95
82.54
93
91.39
96
92.32
96
93.48
96
65.75
95
90.26
96
59.09
94
74.29
94
74.25
94
61.32
94
MEDIAN_ROBtwo views97.56
96
99.08
97
98.39
97
100.00
97
99.99
97
100.00
99
100.00
99
99.49
99
98.31
97
100.00
99
99.95
99
97.97
99
96.05
97
99.97
97
99.41
97
98.43
97
98.53
97
98.66
97
99.88
98
97.95
97
93.38
97
96.45
97
100.00
96
97.23
97
86.96
97
92.33
97
89.80
97
95.90
98
AVERAGE_ROBtwo views98.95
97
100.00
100
99.79
98
100.00
97
100.00
98
100.00
99
100.00
99
100.00
100
100.00
98
100.00
99
100.00
100
100.00
102
100.00
100
100.00
98
99.77
98
99.85
98
100.00
100
100.00
100
100.00
99
100.00
100
98.12
98
99.51
98
100.00
96
100.00
98
93.43
98
93.75
98
94.35
98
93.16
96
DGTPSM_ROBtwo views99.52
98
99.77
98
100.00
99
100.00
97
100.00
98
100.00
99
100.00
99
99.17
97
100.00
98
99.31
97
99.64
98
91.36
97
99.13
98
100.00
98
99.94
99
100.00
99
99.82
98
99.99
98
99.53
96
99.99
98
99.80
100
99.66
99
100.00
96
100.00
98
100.00
99
100.00
99
100.00
99
100.00
99
DPSMNet_ROBtwo views99.53
99
99.78
99
100.00
99
100.00
97
100.00
98
100.00
99
100.00
99
99.17
97
100.00
98
99.31
97
99.63
97
91.42
98
99.14
99
100.00
98
99.94
99
100.00
99
99.82
98
99.99
98
99.62
97
99.99
98
99.79
99
99.66
99
100.00
96
100.00
98
100.00
99
100.00
99
100.00
99
100.00
99
DPSMtwo views99.92
100
100.00
100
100.00
99
100.00
97
100.00
98
98.97
97
98.99
97
100.00
100
100.00
98
100.00
99
100.00
100
99.90
100
100.00
100
100.00
98
100.00
101
100.00
99
100.00
100
100.00
100
100.00
99
100.00
100
100.00
101
100.00
101
100.00
96
100.00
98
100.00
99
100.00
99
100.00
99
100.00
99
DPSM_ROBtwo views99.92
100
100.00
100
100.00
99
100.00
97
100.00
98
98.97
97
98.99
97
100.00
100
100.00
98
100.00
99
100.00
100
99.90
100
100.00
100
100.00
98
100.00
101
100.00
99
100.00
100
100.00
100
100.00
99
100.00
100
100.00
101
100.00
101
100.00
96
100.00
98
100.00
99
100.00
99
100.00
99
100.00
99
STTRV1_RVCtwo views2.31
49
1.06
32
5.61
66
1.39
35
5.42
62
2.59
53
3.68
51
3.77
55
4.47
71
4.40
49
16.44
61
20.14
69
1.02
61
9.83
79
3.26
50
18.06
87
0.81
33
11.45
82
18.05
89
6.27
63
4.23
65
6.12
57
7.90
57
11.99
63
28.90
79