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 bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysorted bysort bysort by
HSMtwo views4.00
14
0.79
12
3.16
7
1.59
12
2.17
43
6.77
24
1.11
16
12.28
29
6.35
4
6.75
13
8.11
7
13.90
34
5.37
31
8.85
26
2.71
8
0.00
1
0.00
1
0.00
1
0.00
1
0.02
2
0.02
2
DISCOtwo views6.28
38
0.57
6
5.78
32
3.43
53
1.17
28
11.22
43
3.39
28
12.14
27
16.16
44
6.52
11
11.22
21
16.96
41
6.32
35
19.51
58
10.74
49
0.00
1
0.00
1
0.00
1
0.00
1
0.35
47
0.11
23
NCCL2two views5.88
32
1.59
45
5.44
28
1.87
19
0.92
22
9.55
34
11.55
61
12.11
26
9.94
26
9.67
26
8.85
11
22.28
51
7.41
41
8.78
24
7.17
32
0.01
14
0.00
1
0.03
45
0.00
1
0.13
25
0.23
36
pmcnntwo views7.72
43
1.27
38
9.42
53
2.91
44
3.14
48
9.44
33
6.23
39
12.56
33
16.51
46
14.53
46
24.08
48
27.44
56
8.49
44
9.32
29
8.44
38
0.06
34
0.08
46
0.00
1
0.00
1
0.30
43
0.15
30
SGM-Foresttwo views4.96
24
0.32
1
2.84
4
1.21
3
0.64
10
10.23
40
6.64
44
11.55
22
10.98
30
10.94
33
13.59
34
11.65
29
4.30
22
8.94
27
4.63
24
0.11
39
0.04
35
0.00
1
0.00
1
0.05
10
0.46
47
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
MSMD_ROBtwo views9.28
50
1.09
28
4.65
20
1.58
11
0.39
5
16.52
53
4.41
33
13.60
36
14.87
39
22.34
52
39.89
62
25.67
53
20.71
58
12.42
42
6.98
30
0.34
48
0.03
31
0.00
1
0.00
1
0.05
10
0.09
17
MDST_ROBtwo views8.37
47
0.32
1
9.03
51
4.18
56
2.42
45
26.86
63
6.14
38
19.36
55
13.52
35
27.09
57
22.75
47
9.47
21
4.74
25
15.06
52
6.34
27
0.02
19
0.02
27
0.00
1
0.00
1
0.02
2
0.13
28
MLCVtwo views3.44
7
0.88
15
5.60
29
1.39
7
0.25
2
4.36
10
0.33
5
7.25
2
7.28
7
9.17
24
12.24
28
5.09
4
2.47
8
9.15
28
3.23
12
0.00
1
0.00
1
0.00
1
0.00
1
0.10
17
0.02
2
iResNetv2_ROBtwo views4.28
20
1.43
42
7.17
43
2.91
44
1.26
29
4.36
10
1.62
21
13.64
37
10.25
28
9.83
27
11.41
24
7.68
14
4.00
18
7.75
8
1.85
2
0.00
1
0.00
1
0.00
1
0.00
1
0.37
49
0.09
17
CBMVpermissivetwo views5.35
28
0.91
17
3.67
11
1.62
13
0.44
8
10.09
38
7.19
48
12.49
31
12.33
33
12.22
36
14.69
39
10.93
26
6.48
36
8.51
20
4.96
25
0.02
19
0.15
51
0.00
1
0.00
1
0.17
30
0.17
32
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
CBMV_ROBtwo views4.14
16
0.52
5
3.14
6
1.30
5
0.77
16
6.92
25
1.97
24
10.11
12
9.58
21
8.92
23
14.20
38
7.12
13
5.90
34
8.65
22
3.50
18
0.01
14
0.05
37
0.00
1
0.00
1
0.04
8
0.09
17
DLCB_ROBtwo views4.51
22
0.91
17
3.78
12
2.19
30
1.07
27
6.28
21
3.09
25
9.78
9
7.72
10
10.65
30
12.97
32
13.91
35
3.71
15
8.72
23
5.30
26
0.00
1
0.00
1
0.00
1
0.00
1
0.03
7
0.10
22
MeshStereopermissivetwo views11.52
56
1.52
44
4.55
19
1.89
20
1.46
32
19.87
58
5.11
35
20.66
59
15.91
43
32.67
62
34.51
60
39.34
65
21.15
59
18.74
57
12.10
53
0.11
39
0.06
41
0.01
35
0.00
1
0.45
51
0.22
35
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
iResNet_ROBtwo views4.23
19
1.02
22
4.90
21
2.18
29
0.93
23
2.92
3
0.37
8
15.10
44
16.91
47
7.89
19
10.51
16
7.03
10
3.07
13
8.16
18
3.46
17
0.01
14
0.00
1
0.00
1
0.00
1
0.10
17
0.02
2
AdaStereotwo views3.09
3
0.58
7
3.04
5
2.84
43
0.48
9
4.08
8
1.29
19
12.16
28
7.77
12
6.03
7
9.62
13
5.79
6
1.53
3
4.56
2
1.93
3
0.00
1
0.00
1
0.00
1
0.00
1
0.02
2
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
HITNettwo views2.79
2
0.77
11
4.02
15
2.03
26
0.11
1
5.58
19
0.59
10
9.24
7
5.15
3
6.42
10
7.26
4
3.66
2
2.92
12
4.07
1
3.87
20
0.00
1
0.00
1
0.00
1
0.00
1
0.06
14
0.02
2
Vladimir Tankovich, Christian Häne, Sean Fanello, Yinda Zhang, Shahram Izadi, Sofien Bouaziz: HITNet: Hierarchical Iterative Tile Refinement Network for Real-time Stereo Matching.
NOSS_ROBtwo views3.30
4
0.46
4
2.62
1
2.08
27
1.01
25
5.60
20
0.74
14
10.37
14
11.48
31
5.15
4
8.43
9
5.67
5
1.73
4
7.97
12
2.34
6
0.02
19
0.06
41
0.00
1
0.00
1
0.07
15
0.14
29
AANet_RVCtwo views5.01
26
1.74
48
6.38
39
1.96
25
1.29
31
2.26
1
1.69
22
10.07
11
18.53
50
7.88
18
18.15
46
8.49
19
2.70
10
10.59
36
7.04
31
0.96
59
0.15
51
0.02
40
0.00
1
0.13
25
0.12
25
ccstwo views3.37
6
1.16
33
3.89
14
2.94
46
0.78
17
4.78
15
0.33
5
9.00
5
7.77
12
5.90
6
10.84
18
7.74
15
2.31
7
7.76
9
1.98
4
0.00
1
0.00
1
0.00
1
0.00
1
0.16
28
0.06
11
HSM-Net_RVCpermissivetwo views4.20
18
0.32
1
2.76
3
0.63
1
0.69
12
6.95
26
1.69
22
11.96
24
8.36
18
8.83
22
12.17
26
15.18
38
4.21
21
6.91
4
3.30
13
0.02
19
0.02
27
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
RYNettwo views6.34
39
0.89
16
5.88
33
1.41
8
4.48
58
15.97
52
4.18
31
13.41
35
16.49
45
10.81
32
7.00
3
14.33
37
8.72
45
9.43
30
13.71
58
0.00
1
0.01
22
0.00
1
0.00
1
0.02
2
0.07
13
PA-Nettwo views4.98
25
1.47
43
7.42
45
2.40
34
2.14
42
8.73
30
3.64
29
12.42
30
13.11
34
7.03
14
7.57
5
7.88
17
6.52
37
10.16
33
7.82
33
0.02
19
0.03
31
0.00
1
0.00
1
0.11
21
1.07
56
Zhibo Rao, Mingyi He, Yuchao Dai, Zhelun Shen: Patch Attention Network with Generative Adversarial Model for Semi-Supervised Binocular Disparity Prediction.
DN-CSS_ROBtwo views2.69
1
1.40
41
5.34
27
2.31
33
0.75
14
3.14
4
0.06
1
6.11
1
3.87
1
5.34
5
12.18
27
2.34
1
1.22
1
7.84
10
1.48
1
0.03
28
0.00
1
0.00
1
0.00
1
0.35
47
0.03
7
PWC_ROBbinarytwo views8.24
46
3.13
58
12.74
58
2.43
35
4.43
55
7.51
27
1.22
17
16.63
49
19.24
53
16.08
48
28.29
51
13.99
36
10.16
49
13.63
48
14.06
59
0.42
52
0.00
1
0.05
50
0.00
1
0.59
52
0.27
42
PWCDC_ROBbinarytwo views7.92
45
3.17
60
7.48
46
5.73
60
4.40
54
10.45
41
0.35
7
14.52
41
28.19
59
10.36
29
31.27
55
7.04
11
9.14
46
13.22
45
8.78
41
2.74
62
0.02
27
0.00
1
0.00
1
1.31
57
0.17
32
TDLMtwo views4.11
15
1.11
30
3.54
9
1.62
13
1.04
26
3.91
6
7.41
49
10.60
18
10.67
29
6.38
9
12.59
30
5.95
7
4.77
26
8.79
25
3.04
10
0.58
53
0.00
1
0.01
35
0.00
1
0.19
32
0.12
25
CFNettwo views3.72
11
1.10
29
5.03
22
2.49
36
1.59
35
4.90
16
0.22
4
11.38
21
9.88
25
4.80
2
11.25
22
6.44
8
3.68
14
8.33
19
3.00
9
0.00
1
0.00
1
0.00
1
0.00
1
0.22
38
0.07
13
CFNet_RVCtwo views3.31
5
0.94
20
2.69
2
1.50
10
2.38
44
2.81
2
0.68
12
8.35
4
7.43
8
4.45
1
9.94
14
10.20
24
4.60
23
6.49
3
3.41
15
0.00
1
0.00
1
0.03
45
0.00
1
0.22
38
0.03
7
ccs_robtwo views3.63
9
1.12
31
4.42
18
2.52
37
0.91
21
5.50
18
0.21
3
10.11
12
9.11
20
6.55
12
11.28
23
8.32
18
2.55
9
7.66
7
2.01
5
0.00
1
0.00
1
0.00
1
0.00
1
0.20
33
0.08
15
iResNettwo views3.68
10
0.91
17
7.94
48
2.97
47
0.34
4
4.44
14
0.48
9
7.70
3
9.74
23
7.72
16
12.74
31
4.03
3
2.87
11
8.05
14
3.37
14
0.02
19
0.01
22
0.00
1
0.00
1
0.10
17
0.09
17
WCMA_ROBtwo views9.21
49
0.87
14
7.37
44
2.54
38
2.13
41
13.59
49
5.80
37
11.64
23
14.01
36
24.43
55
32.99
57
27.09
55
18.02
52
12.51
43
9.85
47
0.81
56
0.07
44
0.01
35
0.01
31
0.16
28
0.23
36
DeepPruner_ROBtwo views3.52
8
1.14
32
4.06
16
1.12
2
1.65
36
3.65
5
0.83
15
13.96
38
4.47
2
7.80
17
10.84
18
7.05
12
2.16
6
8.14
17
3.08
11
0.07
36
0.03
31
0.00
1
0.01
31
0.32
44
0.06
11
CVANet_RVCtwo views4.16
17
1.16
33
3.60
10
1.94
24
1.46
32
3.92
7
4.68
34
10.89
20
8.34
17
7.58
15
10.84
18
10.27
25
6.62
38
8.56
21
2.69
7
0.39
50
0.00
1
0.00
1
0.01
31
0.21
37
0.09
17
SANettwo views10.64
55
1.86
52
10.91
55
1.76
16
0.71
13
14.62
51
9.23
52
19.18
53
37.14
65
19.22
50
27.96
50
25.86
54
19.11
56
13.02
44
10.63
48
0.08
37
0.06
41
0.03
45
0.02
34
0.62
53
0.81
54
ETE_ROBtwo views5.80
30
1.77
50
6.33
38
1.44
9
0.78
17
6.43
23
6.90
45
12.53
32
8.08
14
12.93
43
14.89
40
21.13
50
5.87
33
9.83
31
6.57
29
0.04
29
0.01
22
0.00
1
0.02
34
0.08
16
0.33
44
LALA_ROBtwo views6.58
41
1.80
51
6.25
36
1.26
4
0.94
24
10.08
37
9.02
50
16.00
47
11.51
32
12.74
41
13.02
33
24.77
52
5.25
29
10.56
35
8.02
34
0.04
29
0.05
37
0.00
1
0.02
34
0.10
17
0.25
39
GANetREF_RVCpermissivetwo views6.56
40
2.89
57
7.58
47
3.41
52
0.40
6
12.96
47
9.58
54
15.09
42
17.25
48
10.33
28
10.62
17
12.27
32
8.16
43
12.21
40
4.53
23
0.41
51
0.00
1
0.00
1
0.02
34
3.12
62
0.39
45
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
GANettwo views6.22
37
1.07
25
4.07
17
2.27
31
0.89
19
9.19
31
9.52
53
12.02
25
8.13
15
10.72
31
29.09
53
13.86
33
7.52
42
11.00
37
4.39
22
0.36
49
0.00
1
0.02
40
0.02
34
0.12
23
0.08
15
CC-Net-ROBtwo views3.84
12
1.07
25
5.23
25
2.65
40
2.96
47
4.22
9
0.69
13
10.43
15
7.72
10
8.78
21
8.29
8
9.61
22
4.02
19
7.16
6
3.65
19
0.13
41
0.03
31
0.02
40
0.03
39
0.05
10
0.03
7
DANettwo views6.02
34
1.23
36
8.45
50
3.86
55
3.94
53
7.64
28
1.34
20
9.51
8
7.00
5
13.39
44
15.53
43
15.99
39
7.02
40
12.14
39
12.37
54
0.19
44
0.12
49
0.02
40
0.03
39
0.13
25
0.56
49
StereoDRNettwo views5.59
29
1.75
49
6.80
41
3.12
49
4.45
56
10.61
42
4.35
32
18.80
51
9.73
22
12.22
36
6.87
1
11.44
28
4.65
24
8.09
16
8.26
36
0.02
19
0.11
48
0.00
1
0.03
39
0.20
33
0.28
43
FC-DCNNcopylefttwo views10.41
54
0.74
10
5.05
23
1.74
15
1.53
34
17.47
55
5.18
36
17.42
50
17.59
49
28.46
58
33.95
59
33.48
60
20.55
57
15.46
54
9.31
44
0.06
34
0.01
22
0.00
1
0.03
39
0.05
10
0.11
23
XPNet_ROBtwo views6.03
35
1.22
35
5.61
30
2.56
39
0.90
20
6.32
22
7.07
46
12.92
34
8.30
16
14.76
47
15.13
42
19.84
47
6.66
39
10.36
34
8.58
39
0.02
19
0.04
35
0.00
1
0.03
39
0.11
21
0.24
38
NLCA_NET_v2_RVCtwo views3.84
12
1.06
24
5.23
25
2.72
42
3.27
49
4.36
10
0.61
11
10.71
19
7.56
9
8.75
20
7.89
6
9.86
23
3.90
17
7.15
5
3.44
16
0.14
42
0.02
27
0.02
40
0.03
39
0.04
8
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.
DRN-Testtwo views5.87
31
0.98
21
5.89
34
2.69
41
3.65
52
12.37
44
3.35
27
20.07
58
10.20
27
11.93
35
12.31
29
11.06
27
5.31
30
7.89
11
9.05
42
0.04
29
0.05
37
0.04
49
0.04
45
0.18
31
0.25
39
CSANtwo views7.62
42
1.60
46
6.56
40
1.83
17
0.66
11
12.40
45
10.52
60
14.45
40
21.32
54
14.19
45
15.98
44
17.84
42
13.02
51
12.32
41
8.38
37
0.09
38
0.07
44
0.03
45
0.04
45
0.33
45
0.67
52
StereoDRNet-Refinedtwo views4.46
21
0.62
9
3.80
13
1.92
22
0.40
6
9.35
32
0.15
2
10.02
10
8.83
19
12.69
40
11.62
25
9.34
20
3.87
16
8.06
15
8.02
34
0.00
1
0.00
1
0.01
35
0.05
47
0.20
33
0.26
41
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
NVstereo2Dtwo views4.51
22
0.82
13
6.86
42
3.28
51
3.38
51
8.16
29
3.13
26
10.51
16
15.15
40
4.90
3
6.89
2
7.87
16
4.78
27
9.88
32
3.91
21
0.01
14
0.00
1
0.00
1
0.06
48
0.02
2
0.58
50
Anonymous Stereotwo views6.16
36
3.15
59
23.75
65
2.97
47
2.48
46
4.39
13
13.30
63
9.21
6
9.86
24
9.56
25
8.76
10
6.79
9
1.99
5
13.50
46
13.04
57
0.01
14
0.05
37
0.00
1
0.06
48
0.22
38
0.19
34
PDISCO_ROBtwo views9.62
51
1.99
54
11.51
56
9.88
65
9.61
64
21.48
59
3.83
30
19.33
54
28.49
60
11.27
34
14.17
37
19.92
48
5.02
28
16.35
55
9.18
43
5.28
64
0.41
56
0.14
55
0.09
50
2.05
61
2.36
61
PSMNet_ROBtwo views5.02
27
1.63
47
6.03
35
1.90
21
1.83
40
9.57
35
6.35
42
15.58
46
7.23
6
6.15
8
10.48
15
12.22
31
4.16
20
8.02
13
8.71
40
0.02
19
0.01
22
0.01
35
0.10
51
0.20
33
0.12
25
NaN_ROBtwo views6.00
33
1.24
37
6.29
37
1.34
6
1.68
38
9.60
36
10.31
58
15.09
42
15.79
42
12.62
39
8.95
12
11.67
30
5.83
32
11.78
38
6.41
28
0.05
32
0.13
50
0.08
51
0.20
52
0.22
38
0.79
53
LE_ROBtwo views16.73
64
1.28
40
11.61
57
3.72
54
1.65
36
16.67
54
9.17
51
14.39
39
55.91
68
63.81
68
40.86
65
35.94
62
37.73
68
14.24
50
26.87
67
0.05
32
0.10
47
0.13
54
0.22
53
0.12
23
0.15
30
MFMNet_retwo views13.29
57
8.60
66
18.29
62
9.75
64
7.25
62
19.65
57
14.84
64
20.71
60
30.72
62
23.03
53
28.77
52
18.85
45
26.09
64
13.55
47
9.82
45
2.44
61
1.35
63
0.34
59
0.23
54
4.78
64
6.69
63
SGM_RVCbinarytwo views10.08
52
0.60
8
3.42
8
2.30
32
0.32
3
19.41
56
6.33
41
18.95
52
14.64
37
25.14
56
24.32
49
33.34
59
18.79
55
19.86
59
12.55
56
0.25
46
0.26
54
0.22
56
0.24
55
0.34
46
0.40
46
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
DPSNettwo views10.14
53
1.88
53
16.82
61
1.85
18
1.73
39
24.84
61
17.20
67
19.92
57
27.41
58
12.23
38
13.62
35
16.52
40
18.35
53
14.42
51
12.50
55
0.78
55
0.54
59
0.08
51
0.25
56
1.18
56
0.59
51
LSMtwo views14.01
58
5.95
62
33.49
66
6.78
62
43.61
68
10.22
39
9.98
57
15.16
45
22.93
55
23.07
54
32.34
56
18.52
44
12.67
50
15.45
53
11.10
52
0.16
43
0.51
58
0.09
53
0.32
57
1.08
55
16.85
68
SGM-ForestMtwo views16.99
65
1.08
27
5.74
31
2.12
28
0.75
14
31.63
66
12.21
62
27.80
66
32.25
63
37.88
63
39.99
63
52.96
68
35.20
67
33.60
67
24.47
65
0.26
47
0.39
55
0.31
58
0.39
58
0.26
42
0.53
48
ELAS_RVCcopylefttwo views16.54
62
2.26
56
10.09
54
5.50
59
4.46
57
28.28
64
16.72
66
25.55
64
33.54
64
40.19
64
40.30
64
36.68
63
30.03
65
29.40
65
20.61
62
0.98
60
1.21
61
0.86
61
0.70
59
1.39
58
2.16
59
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
FBW_ROBtwo views8.50
48
1.03
23
7.98
49
1.93
23
1.28
30
13.10
48
6.23
39
22.50
61
18.98
51
18.82
49
14.91
41
19.06
46
10.04
48
18.41
56
9.83
46
0.62
54
0.22
53
1.82
63
0.82
60
0.99
54
1.36
58
ELAScopylefttwo views16.72
63
2.14
55
9.23
52
4.92
58
4.53
59
32.66
67
15.11
65
27.40
65
28.68
61
40.27
65
44.90
66
38.33
64
30.50
66
26.44
64
21.94
64
0.88
57
1.23
62
0.67
60
0.89
61
1.49
59
2.18
60
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
PASMtwo views7.90
44
4.22
61
21.97
64
3.25
50
3.29
50
5.39
17
6.57
43
10.57
17
19.09
52
12.77
42
13.92
36
18.11
43
9.51
47
13.79
49
10.77
51
0.19
44
0.45
57
0.29
57
1.08
62
1.49
59
1.19
57
SPS-STEREOcopylefttwo views15.04
59
6.23
63
13.21
60
11.34
66
11.65
66
23.30
60
7.15
47
24.16
63
15.65
41
31.78
61
29.19
54
31.62
57
21.32
60
24.62
62
19.50
61
7.59
66
4.19
67
3.22
64
1.48
63
6.99
66
6.54
62
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
SGM+DAISYtwo views15.62
60
7.26
65
19.28
63
8.94
63
10.11
65
26.25
62
10.49
59
19.36
55
14.65
38
30.64
60
33.59
58
33.00
58
22.32
61
24.96
63
16.42
60
7.90
67
6.25
68
4.51
65
3.37
64
5.86
65
7.20
64
PWCKtwo views30.53
68
44.32
68
47.25
69
29.76
68
7.23
61
40.78
68
27.10
68
44.73
68
44.32
67
47.31
67
36.37
61
47.16
66
26.05
63
41.26
68
31.87
68
21.83
68
4.03
66
29.50
67
4.67
65
27.17
68
7.80
65
DispFullNettwo views17.47
66
26.01
67
33.98
67
22.58
67
20.86
67
13.84
50
1.28
18
16.50
48
26.27
57
19.97
51
17.17
45
20.52
49
18.49
54
22.86
61
10.76
50
5.13
63
2.83
64
30.72
68
7.72
66
20.86
67
11.01
67
MANEtwo views19.47
67
1.27
38
5.07
24
4.69
57
5.55
60
30.49
65
9.94
56
34.01
67
37.27
66
44.13
66
51.57
69
52.51
67
40.41
69
33.58
66
24.81
66
0.89
58
0.86
60
1.11
62
9.72
67
0.38
50
1.06
55
NVStereoNet_ROBtwo views16.04
61
6.75
64
12.90
59
6.37
61
7.42
63
12.89
46
9.74
55
22.78
62
25.12
56
30.32
59
46.19
67
34.37
61
25.38
62
21.48
60
21.38
63
5.94
65
3.10
65
6.07
66
10.09
68
4.01
63
8.54
66
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
DPSimNet_ROBtwo views53.45
69
64.73
69
44.39
68
53.97
69
45.39
69
53.66
69
54.83
69
55.15
69
57.87
69
64.16
69
50.83
68
63.40
69
53.34
70
46.45
69
65.81
69
63.13
69
26.54
69
57.94
69
51.11
69
45.52
69
50.69
69
MEDIAN_ROBtwo views98.41
70
99.70
70
99.30
71
97.09
70
97.02
70
96.89
70
95.77
71
97.66
70
97.28
70
98.79
72
98.94
70
99.18
70
98.14
71
96.89
70
96.88
70
99.96
72
99.16
70
100.00
70
99.99
70
99.69
70
99.88
70
DPSMtwo views99.95
74
100.00
72
100.00
74
99.76
71
100.00
71
100.00
72
100.00
72
100.00
73
100.00
71
100.00
73
100.00
71
100.00
73
100.00
72
100.00
73
100.00
74
99.21
70
100.00
71
100.00
70
100.00
71
99.99
71
99.95
71
DPSM_ROBtwo views99.95
74
100.00
72
100.00
74
99.76
71
100.00
71
100.00
72
100.00
72
100.00
73
100.00
71
100.00
73
100.00
71
100.00
73
100.00
72
100.00
73
100.00
74
99.21
70
100.00
71
100.00
70
100.00
71
99.99
71
99.95
71
DGTPSM_ROBtwo views99.90
72
100.00
72
99.99
72
99.99
73
100.00
71
100.00
72
100.00
72
99.97
71
100.00
71
98.35
70
100.00
71
99.84
71
100.00
72
99.98
71
99.99
71
99.99
73
100.00
71
100.00
70
100.00
71
100.00
73
100.00
74
AVERAGE_ROBtwo views99.62
71
99.95
71
98.81
70
100.00
75
100.00
71
98.08
71
95.47
70
100.00
73
100.00
71
100.00
73
100.00
71
100.00
73
100.00
72
100.00
73
99.99
71
100.00
74
100.00
71
100.00
70
100.00
71
100.00
73
100.00
74
DPSMNet_ROBtwo views99.91
73
100.00
72
99.99
72
99.99
73
100.00
71
100.00
72
100.00
72
99.98
72
100.00
71
98.35
70
100.00
71
99.84
71
100.00
72
99.98
71
99.99
71
100.00
74
100.00
71
100.00
70
100.00
71
100.00
73
100.00
74
LSM0two views100.00
76
100.00
72
100.00
74
100.00
75
100.00
71
100.00
72
100.00
72
100.00
73
100.00
71
100.00
73
100.00
71
100.00
73
100.00
72
100.00
73
100.00
74
100.00
74
100.00
71
100.00
70
100.00
71
100.00
73
99.99
73
MSMDNettwo views1.26
2