Low-res two-view results

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):

bad 0.5, 1.0, 2.0, 4.0 [%]: Fraction of pixels with errors larger than the given number of disparities.
Average error [px]: The per-pixel average disparity error.
Root mean square error [px]: The per-pixel root mean square disparity error.
50%, 90%, 95%, 99% error quantile [px]: The highest disparity error within the given percentage of best pixels (for 50%, this is the median error).
Time [s]: The runtime of the method.

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	all	lakes. 1l	lakes. 1s	sand box 1l	sand box 1s	stora. room 1l	stora. room 1s	stora. room 2l	stora. room 2s	stora. room 2 1l	stora. room 2 1s	stora. room 2 2l	stora. room 2 2s	stora. room 3l	stora. room 3s	tunnel 1l	tunnel 1s	tunnel 2l	tunnel 2s	tunnel 3l	tunnel 3s

DN-CSS_ROB	2.69 1	1.40 36	5.34 24	2.31 30	0.75 12	3.14 3	0.06 1	6.11 1	3.87 1	5.34 5	12.18 23	2.34 1	1.22 1	7.84 4	1.48 1	0.03 26	0.00 1	0.00 1	0.00 1	0.35 43	0.03 9

AdaStereo	3.41 5	0.58 6	5.33 23	3.52 48	0.85 16	4.08 8	1.29 18	12.16 24	7.77 11	6.57 11	9.62 14	7.82 14	1.53 2	4.85 2	2.20 4	0.00 1	0.02 26	0.00 1	0.00 1	0.02 1	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
NOSS_ROB	3.30 3	0.46 3	2.62 1	2.08 23	1.01 23	5.60 18	0.74 13	10.37 13	11.48 29	5.15 4	8.43 7	5.67 5	1.73 3	7.97 6	2.34 5	0.02 19	0.06 37	0.00 1	0.00 1	0.07 12	0.14 26

Anonymous Stereo	6.16 32	3.15 54	23.75 60	2.97 39	2.48 44	4.39 11	13.30 58	9.21 5	9.86 24	9.56 22	8.76 8	6.79 6	1.99 4	13.50 42	13.04 54	0.01 14	0.05 33	0.00 1	0.06 44	0.22 36	0.19 31

DeepPruner_ROB	3.52 7	1.14 27	4.06 15	1.12 1	1.65 33	3.65 5	0.83 14	13.96 33	4.47 2	7.80 16	10.84 18	7.05 9	2.16 5	8.14 11	3.08 9	0.07 34	0.03 29	0.00 1	0.01 29	0.32 40	0.06 12

MLCV	3.44 6	0.88 14	5.60 26	1.39 6	0.25 2	4.36 9	0.33 3	7.25 2	7.28 8	9.17 20	12.24 24	5.09 4	2.47 6	9.15 24	3.23 10	0.00 1	0.00 1	0.00 1	0.00 1	0.10 14	0.02 2

AANet_RVC	5.01 22	1.74 43	6.38 35	1.96 21	1.29 28	2.26 1	1.69 21	10.07 10	18.53 46	7.88 17	18.15 41	8.49 19	2.70 7	10.59 32	7.04 29	0.96 54	0.15 47	0.02 38	0.00 1	0.13 22	0.12 23

iResNet	3.68 10	0.91 15	7.94 43	2.97 39	0.34 4	4.44 12	0.48 9	7.70 3	9.74 22	7.72 14	12.74 26	4.03 3	2.87 8	8.05 8	3.37 11	0.02 19	0.01 22	0.00 1	0.00 1	0.10 14	0.09 15

ccs	3.64 9	1.28 34	2.95 3	3.00 42	1.87 39	5.46 16	0.63 12	10.87 17	9.22 18	5.73 6	8.91 11	7.36 11	2.87 8	8.74 20	3.63 14	0.00 1	0.00 1	0.00 1	0.00 1	0.19 28	0.02 2

HITNet	2.79 2	0.77 10	4.02 14	2.03 22	0.11 1	5.58 17	0.59 10	9.24 6	5.15 3	6.42 9	7.26 3	3.66 2	2.92 10	4.07 1	3.87 15	0.00 1	0.00 1	0.00 1	0.00 1	0.06 11	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.
CPSMnet	3.38 4	1.21 29	3.40 7	2.53 32	1.57 32	4.61 13	0.47 8	10.82 16	7.94 12	4.77 2	8.01 4	7.87 15	3.00 11	8.38 15	2.86 8	0.00 1	0.00 1	0.00 1	0.00 1	0.20 29	0.01 1

ccs_rob	3.53 8	1.01 20	2.95 3	3.81 50	1.80 37	3.33 4	0.46 7	12.77 29	7.16 6	4.75 1	8.10 5	7.38 12	3.01 12	9.86 28	3.99 17	0.00 1	0.00 1	0.02 38	0.00 1	0.12 19	0.06 12

iResNet_ROB	4.23 16	1.02 21	4.90 20	2.18 25	0.93 20	2.92 2	0.37 6	15.10 40	16.91 43	7.89 18	10.51 16	7.03 7	3.07 13	8.16 12	3.46 12	0.01 14	0.00 1	0.00 1	0.00 1	0.10 14	0.02 2

CC-Net	3.68 10	0.92 18	4.90 20	2.98 41	2.97 45	3.77 6	0.60 11	8.94 4	7.68 9	9.31 21	9.45 13	8.30 18	3.41 14	8.25 14	2.12 3	0.00 1	0.00 1	0.00 1	0.00 1	0.02 1	0.02 2

DLCB_ROB	4.51 19	0.91 15	3.78 11	2.19 26	1.07 24	6.28 19	3.09 23	9.78 8	7.72 10	10.65 27	12.97 27	13.91 32	3.71 15	8.72 19	5.30 22	0.00 1	0.00 1	0.00 1	0.00 1	0.03 6	0.10 20

StereoDRNet-Refined	4.46 18	0.62 8	3.80 12	1.92 18	0.40 6	9.35 29	0.15 2	10.02 9	8.83 17	12.69 36	11.62 22	9.34 20	3.87 16	8.06 9	8.02 31	0.00 1	0.00 1	0.01 34	0.05 43	0.20 29	0.26 38
Rohan Chabra, Julian Straub, Chris Sweeney, Richard Newcombe, Henry Fuchs: StereoDRNet. CVPR
iResNetv2_ROB	4.28 17	1.43 37	7.17 39	2.91 37	1.26 26	4.36 9	1.62 20	13.64 32	10.25 27	9.83 24	11.41 21	7.68 13	4.00 17	7.75 3	1.85 2	0.00 1	0.00 1	0.00 1	0.00 1	0.37 45	0.09 15

PSMNet_ROB	5.02 23	1.63 42	6.03 31	1.90 17	1.83 38	9.57 32	6.35 38	15.58 42	7.23 7	6.15 8	10.48 15	12.22 28	4.16 18	8.02 7	8.71 37	0.02 19	0.01 22	0.01 34	0.10 47	0.20 29	0.12 23

SGM-Forest	4.96 21	0.32 1	2.84 2	1.21 2	0.64 9	10.23 37	6.64 40	11.55 19	10.98 28	10.94 29	13.59 29	11.65 26	4.30 19	8.94 23	4.63 20	0.11 37	0.04 31	0.00 1	0.00 1	0.05 8	0.46 44
Johannes L. Schönberger, Sudipta Sinha, Marc Pollefeys: Learning to Fuse Proposals from Multiple Scanline Optimizations in Semi-Global Matching. ECCV 2018
CPSMnet_ROB	3.84 12	1.07 23	3.81 13	2.87 36	2.40 42	4.79 14	0.33 3	10.29 12	9.55 19	5.97 7	8.89 10	8.24 17	4.40 20	8.23 13	5.71 24	0.00 1	0.00 1	0.00 1	0.00 1	0.20 29	0.03 9

StereoDRNet	5.59 25	1.75 44	6.80 37	3.12 43	4.45 53	10.61 39	4.35 28	18.80 47	9.73 21	12.22 32	6.87 1	11.44 25	4.65 21	8.09 10	8.26 33	0.02 19	0.11 44	0.00 1	0.03 37	0.20 29	0.28 40

MDST_ROB	8.37 43	0.32 1	9.03 46	4.18 52	2.42 43	26.86 59	6.14 34	19.36 51	13.52 32	27.09 53	22.75 42	9.47 21	4.74 22	15.06 48	6.34 25	0.02 19	0.02 26	0.00 1	0.00 1	0.02 1	0.13 25

NVstereo2D	4.51 19	0.82 12	6.86 38	3.28 45	3.38 48	8.16 26	3.13 24	10.51 14	15.15 37	4.90 3	6.89 2	7.87 15	4.78 23	9.88 29	3.91 16	0.01 14	0.00 1	0.00 1	0.06 44	0.02 1	0.58 47

PDISCO_ROB	9.62 47	1.99 49	11.51 51	9.88 60	9.61 59	21.48 55	3.83 27	19.33 50	28.49 56	11.27 30	14.17 32	19.92 44	5.02 24	16.35 51	9.18 40	5.28 59	0.41 52	0.14 51	0.09 46	2.05 56	2.36 56

LALA_ROB	6.58 37	1.80 46	6.25 32	1.26 3	0.94 21	10.08 34	9.02 45	16.00 43	11.51 30	12.74 37	13.02 28	24.77 48	5.25 25	10.56 31	8.02 31	0.04 27	0.05 33	0.00 1	0.02 32	0.10 14	0.25 36

DRN-Test	5.87 27	0.98 19	5.89 30	2.69 35	3.65 49	12.37 41	3.35 25	20.07 54	10.20 26	11.93 31	12.31 25	11.06 24	5.31 26	7.89 5	9.05 39	0.04 27	0.05 33	0.04 45	0.04 41	0.18 27	0.25 36

HSM	4.00 13	0.79 11	3.16 6	1.59 9	2.17 41	6.77 22	1.11 15	12.28 25	6.35 4	6.75 12	8.11 6	13.90 31	5.37 27	8.85 22	2.71 7	0.00 1	0.00 1	0.00 1	0.00 1	0.02 1	0.02 2

AMf	6.26 34	1.53 39	4.57 18	2.19 26	0.98 22	8.95 27	10.34 53	14.36 34	9.76 23	7.77 15	25.50 45	18.16 39	5.82 28	9.54 26	5.38 23	0.17 40	0.00 1	0.00 1	0.00 1	0.21 34	0.05 11

NaN_ROB	6.00 29	1.24 32	6.29 33	1.34 5	1.68 35	9.60 33	10.31 52	15.09 38	15.79 39	12.62 35	8.95 12	11.67 27	5.83 29	11.78 34	6.41 26	0.05 30	0.13 46	0.08 47	0.20 48	0.22 36	0.79 50

ETE_ROB	5.80 26	1.77 45	6.33 34	1.44 7	0.78 15	6.43 21	6.90 41	12.53 27	8.08 13	12.93 39	14.89 35	21.13 46	5.87 30	9.83 27	6.57 27	0.04 27	0.01 22	0.00 1	0.02 32	0.08 13	0.33 41

CBMV_ROB	4.14 14	0.52 4	3.14 5	1.30 4	0.77 14	6.92 23	1.97 22	10.11 11	9.58 20	8.92 19	14.20 33	7.12 10	5.90 31	8.65 18	3.50 13	0.01 14	0.05 33	0.00 1	0.00 1	0.04 7	0.09 15

DISCO	6.28 35	0.57 5	5.78 29	3.43 47	1.17 25	11.22 40	3.39 26	12.14 23	16.16 41	6.52 10	11.22 20	16.96 36	6.32 32	19.51 54	10.74 46	0.00 1	0.00 1	0.00 1	0.00 1	0.35 43	0.11 21

CBMV	5.35 24	0.91 15	3.67 10	1.62 10	0.44 8	10.09 35	7.19 44	12.49 26	12.33 31	12.22 32	14.69 34	10.93 23	6.48 33	8.51 16	4.96 21	0.02 19	0.15 47	0.00 1	0.00 1	0.17 26	0.17 29
Konstantinos Batsos, Changjiang Cai, Philippos Mordohai: CBMV: A Coalesced Bidirectional Matching Volume for Disparity Estimation. Computer Vision and Pattern Recognition (CVPR) 2018
CVANet	4.16 15	1.16 28	3.60 9	1.94 20	1.46 29	3.92 7	4.68 30	10.89 18	8.34 16	7.58 13	10.84 18	10.27 22	6.62 34	8.56 17	2.69 6	0.39 47	0.00 1	0.00 1	0.01 29	0.21 34	0.09 15

XPNet_ROB	6.03 31	1.22 30	5.61 27	2.56 34	0.90 18	6.32 20	7.07 42	12.92 30	8.30 15	14.76 43	15.13 37	19.84 43	6.66 35	10.36 30	8.58 36	0.02 19	0.04 31	0.00 1	0.03 37	0.11 18	0.24 35

DANet	6.02 30	1.23 31	8.45 45	3.86 51	3.94 50	7.64 25	1.34 19	9.51 7	7.00 5	13.39 40	15.53 38	15.99 34	7.02 36	12.14 35	12.37 51	0.19 41	0.12 45	0.02 38	0.03 37	0.13 22	0.56 46

NCCL2	5.88 28	1.59 40	5.44 25	1.87 15	0.92 19	9.55 31	11.55 56	12.11 22	9.94 25	9.67 23	8.85 9	22.28 47	7.41 37	8.78 21	7.17 30	0.01 14	0.00 1	0.03 42	0.00 1	0.13 22	0.23 33

GANet	6.22 33	1.07 23	4.07 16	2.27 28	0.89 17	9.19 28	9.52 48	12.02 21	8.13 14	10.72 28	29.09 49	13.86 30	7.52 38	11.00 33	4.39 18	0.36 46	0.00 1	0.02 38	0.02 32	0.12 19	0.08 14

GANetREF_RVC	6.56 36	2.89 52	7.58 42	3.41 46	0.40 6	12.96 44	9.58 49	15.09 38	17.25 44	10.33 25	10.62 17	12.27 29	8.16 39	12.21 36	4.53 19	0.41 48	0.00 1	0.00 1	0.02 32	3.12 57	0.39 42
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
pmcnn	7.72 39	1.27 33	9.42 48	2.91 37	3.14 46	9.44 30	6.23 35	12.56 28	16.51 42	14.53 42	24.08 43	27.44 52	8.49 40	9.32 25	8.44 35	0.06 32	0.08 42	0.00 1	0.00 1	0.30 39	0.15 27

PWCDC_ROB	7.92 41	3.17 55	7.48 41	5.73 55	4.40 51	10.45 38	0.35 5	14.52 37	28.19 55	10.36 26	31.27 51	7.04 8	9.14 41	13.22 41	8.78 38	2.74 57	0.02 26	0.00 1	0.00 1	1.31 52	0.17 29

PASM	7.90 40	4.22 56	21.97 59	3.25 44	3.29 47	5.39 15	6.57 39	10.57 15	19.09 48	12.77 38	13.92 31	18.11 38	9.51 42	13.79 45	10.77 48	0.19 41	0.45 53	0.29 53	1.08 58	1.49 54	1.19 52

FBW_ROB	8.50 44	1.03 22	7.98 44	1.93 19	1.28 27	13.10 45	6.23 35	22.50 57	18.98 47	18.82 45	14.91 36	19.06 42	10.04 43	18.41 52	9.83 43	0.62 50	0.22 49	1.82 58	0.82 56	0.99 49	1.36 53

PWC_ROB	8.24 42	3.13 53	12.74 53	2.43 31	4.43 52	7.51 24	1.22 16	16.63 45	19.24 49	16.08 44	28.29 47	13.99 33	10.16 44	13.63 44	14.06 55	0.42 49	0.00 1	0.05 46	0.00 1	0.59 47	0.27 39

LSM	14.01 54	5.95 57	33.49 61	6.78 57	43.61 63	10.22 36	9.98 51	15.16 41	22.93 51	23.07 50	32.34 52	18.52 40	12.67 45	15.45 49	11.10 49	0.16 39	0.51 54	0.09 49	0.32 53	1.08 50	16.85 63

CSAN	7.62 38	1.60 41	6.56 36	1.83 13	0.66 10	12.40 42	10.52 55	14.45 36	21.32 50	14.19 41	15.98 39	17.84 37	13.02 46	12.32 37	8.38 34	0.09 36	0.07 40	0.03 42	0.04 41	0.33 41	0.67 49

WCMA_ROB	9.21 45	0.87 13	7.37 40	2.54 33	2.13 40	13.59 46	5.80 33	11.64 20	14.01 33	24.43 51	32.99 53	27.09 51	18.02 47	12.51 39	9.85 44	0.81 52	0.07 40	0.01 34	0.01 29	0.16 25	0.23 33

DPSNet	10.14 49	1.88 48	16.82 56	1.85 14	1.73 36	24.84 57	17.20 62	19.92 53	27.41 54	12.23 34	13.62 30	16.52 35	18.35 48	14.42 47	12.50 52	0.78 51	0.54 55	0.08 47	0.25 52	1.18 51	0.59 48

DispFullNet	17.47 62	26.01 62	33.98 62	22.58 62	20.86 62	13.84 47	1.28 17	16.50 44	26.27 53	19.97 47	17.17 40	20.52 45	18.49 49	22.86 57	10.76 47	5.13 58	2.83 59	30.72 63	7.72 62	20.86 62	11.01 62

SGM_RVC	10.08 48	0.60 7	3.42 8	2.30 29	0.32 3	19.41 52	6.33 37	18.95 48	14.64 34	25.14 52	24.32 44	33.34 55	18.79 50	19.86 55	12.55 53	0.25 43	0.26 50	0.22 52	0.24 51	0.34 42	0.40 43
Heiko Hirschmueller: Stereo processing by semiglobal matching and mutual information. TPAMI 2008, Volume 30(2), pp. 328-341
SANet	10.64 51	1.86 47	10.91 50	1.76 12	0.71 11	14.62 48	9.23 47	19.18 49	37.14 61	19.22 46	27.96 46	25.86 50	19.11 51	13.02 40	10.63 45	0.08 35	0.06 37	0.03 42	0.02 32	0.62 48	0.81 51

FC-DCNN	10.41 50	0.74 9	5.05 22	1.74 11	1.53 31	17.47 51	5.18 32	17.42 46	17.59 45	28.46 54	33.95 55	33.48 56	20.55 52	15.46 50	9.31 41	0.06 32	0.01 22	0.00 1	0.03 37	0.05 8	0.11 21

MSMD_ROB	9.28 46	1.09 26	4.65 19	1.58 8	0.39 5	16.52 49	4.41 29	13.60 31	14.87 36	22.34 48	39.89 58	25.67 49	20.71 53	12.42 38	6.98 28	0.34 45	0.03 29	0.00 1	0.00 1	0.05 8	0.09 15

MeshStereo	11.52 52	1.52 38	4.55 17	1.89 16	1.46 29	19.87 54	5.11 31	20.66 55	15.91 40	32.67 58	34.51 56	39.34 61	21.15 54	18.74 53	12.10 50	0.11 37	0.06 37	0.01 34	0.00 1	0.45 46	0.22 32
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
SPS-STEREO	15.04 55	6.23 58	13.21 55	11.34 61	11.65 61	23.30 56	7.15 43	24.16 59	15.65 38	31.78 57	29.19 50	31.62 53	21.32 55	24.62 58	19.50 57	7.59 61	4.19 62	3.22 59	1.48 59	6.99 61	6.54 57
K. Yamaguchi, D. McAllester, R. Urtasun: Efficient Joint Segmentation, Occlusion Labeling, Stereo and Flow Estimation. ECCV 2014
SGM+DAISY	15.62 56	7.26 60	19.28 58	8.94 58	10.11 60	26.25 58	10.49 54	19.36 51	14.65 35	30.64 56	33.59 54	33.00 54	22.32 56	24.96 59	16.42 56	7.90 62	6.25 63	4.51 60	3.37 60	5.86 60	7.20 59

NVStereoNet_ROB	16.04 57	6.75 59	12.90 54	6.37 56	7.42 58	12.89 43	9.74 50	22.78 58	25.12 52	30.32 55	46.19 63	34.37 57	25.38 57	21.48 56	21.38 59	5.94 60	3.10 60	6.07 61	10.09 63	4.01 58	8.54 61
Nikolai Smolyanskiy, Alexey Kamenev, Stan Birchfield: On the Importance of Stereo for Accurate Depth Estimation: An Efficient Semi-Supervised Deep Neural Network Approach. Arxiv
PWCK	30.53 63	44.32 63	47.25 64	29.76 63	7.23 56	40.78 63	27.10 63	44.73 63	44.32 62	47.31 62	36.37 57	47.16 62	26.05 58	41.26 63	31.87 63	21.83 63	4.03 61	29.50 62	4.67 61	27.17 63	7.80 60

MFMNet_re	13.29 53	8.60 61	18.29 57	9.75 59	7.25 57	19.65 53	14.84 59	20.71 56	30.72 58	23.03 49	28.77 48	18.85 41	26.09 59	13.55 43	9.82 42	2.44 56	1.35 58	0.34 55	0.23 50	4.78 59	6.69 58

ELAS_RVC	16.54 58	2.26 51	10.09 49	5.50 54	4.46 54	28.28 60	16.72 61	25.55 60	33.54 60	40.19 60	40.30 60	36.68 59	30.03 60	29.40 61	20.61 58	0.98 55	1.21 56	0.86 57	0.70 55	1.39 53	2.16 54
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
ELAS	16.72 59	2.14 50	9.23 47	4.92 53	4.53 55	32.66 62	15.11 60	27.40 61	28.68 57	40.27 61	44.90 62	38.33 60	30.50 61	26.44 60	21.94 60	0.88 53	1.23 57	0.67 56	0.89 57	1.49 54	2.18 55
A. Geiger, M. Roser, R. Urtasun: Efficient large-scale stereo matching. ACCV 2010
SGM-ForestM	16.99 61	1.08 25	5.74 28	2.12 24	0.75 12	31.63 61	12.21 57	27.80 62	32.25 59	37.88 59	39.99 59	52.96 63	35.20 62	33.60 62	24.47 61	0.26 44	0.39 51	0.31 54	0.39 54	0.26 38	0.53 45

LE_ROB	16.73 60	1.28 34	11.61 52	3.72 49	1.65 33	16.67 50	9.17 46	14.39 35	55.91 63	63.81 63	40.86 61	35.94 58	37.73 63	14.24 46	26.87 62	0.05 30	0.10 43	0.13 50	0.22 49	0.12 19	0.15 27

DPSimNet_ROB	53.45 64	64.73 64	44.39 63	53.97 64	45.39 64	53.66 64	54.83 64	55.15 64	57.87 64	64.16 64	50.83 64	63.40 64	53.34 64	46.45 64	65.81 64	63.13 64	26.54 64	57.94 64	51.11 64	45.52 64	50.69 64

MEDIAN_ROB	98.41 65	99.70 65	99.30 66	97.09 65	97.02 65	96.89 65	95.77 66	97.66 65	97.28 65	98.79 67	98.94 65	99.18 65	98.14 65	96.89 65	96.88 65	99.96 67	99.16 65	100.00 65	99.99 65	99.69 65	99.88 65

AVERAGE_ROB	99.62 66	99.95 66	98.81 65	100.00 70	100.00 66	98.08 66	95.47 65	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	99.99 66	100.00 69	100.00 66	100.00 65	100.00 66	100.00 68	100.00 69

LSM0	100.00 71	100.00 67	100.00 69	100.00 70	100.00 66	100.00 67	100.00 67	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 69	100.00 69	100.00 66	100.00 65	100.00 66	100.00 68	99.99 68

DPSM	99.95 69	100.00 67	100.00 69	99.76 66	100.00 66	100.00 67	100.00 67	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 69	99.21 65	100.00 66	100.00 65	100.00 66	99.99 66	99.95 66

DPSM_ROB	99.95 69	100.00 67	100.00 69	99.76 66	100.00 66	100.00 67	100.00 67	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 66	100.00 68	100.00 69	99.21 65	100.00 66	100.00 65	100.00 66	99.99 66	99.95 66

DGTPSM_ROB	99.90 67	100.00 67	99.99 67	99.99 68	100.00 66	100.00 67	100.00 67	99.97 66	100.00 66	98.35 65	100.00 66	99.84 66	100.00 66	99.98 66	99.99 66	99.99 68	100.00 66	100.00 65	100.00 66	100.00 68	100.00 69

DPSMNet_ROB	99.91 68	100.00 67	99.99 67	99.99 68	100.00 66	100.00 67	100.00 67	99.98 67	100.00 66	98.35 65	100.00 66	99.84 66	100.00 66	99.98 66	99.99 66	100.00 69	100.00 66	100.00 65	100.00 66	100.00 68	100.00 69

Low-res two-view benchmark