Fig. 1. Support window coordinates and weight distribution map with size of 3×3 and standard deviation of 1.5. (a) coordinate distribution map; (b) weight coordinate distribution map; (c) weight distribution map
Fig. 2. Disparity maps before and after improvement by guidance filter algorithm. (a)(c) before improvement; (b)(d) after improvement
Fig. 3. Disparity maps obtained by different matching cost algorithms. (a) CT; (b) MCT; (c) GRD;(d) proposed algorithm
Fig. 4. Disparity maps obtained by different aggregation algorithms. (a) BoxF, R=10.81%; (b) BF, R=8.12%;(c) GF, R=7.85%; (d) MST, R=8.31%; (e) proposed algorithm, R=7.57%
Fig. 5. Comparison of false matching rates of different cost aggregation algorithms in no-occluded regions
Fig. 6. Test results of different stereo matching algorithms on Aloe image pairs. (a) Aloe left image;(b) Aloe right image; (c) GRD, R=10.19%; (d) MCT, R=10.07%; (e) MCT', R=9.74%; (f) proposed algorithm, R=6.74%
Fig. 7. Test results of different stereo matching algorithms on Baby1 image pairs. (a) Baby1 left image; (b) Baby1 right image; (c) GRD, R=12.82%; (d) MCT, R=4.91%; (e) MCT', R=4.53%; (f) proposed algorithm, R=3.75%
Fig. 8. Test results of different stereo matching algorithms on Bowling2 image pairs. (a) Bowling2 left image; (b) Bowling2 right image; (c) GRD, R=12.91%; (d) MCT, R=15.77%; (e) MCT', R=14.21%; (f) proposed algorithm, R=9.04%
Fig. 9. Test results of different stereo matching algorithms on Dolls image pairs. (a) Dolls left image; (b) Dolls right image; (c) GRD, R=7.66%;(d) MCT, R=11.55%; (e) MCT', R=11.87%; (f) proposed algorithm, R=5.77%
Fig. 10. Experimental results of proposed algorithm on Middlebury2.0 image pairs. (a) Testing left image; (b) standard disparity map; (c) disparity map of proposed algorithm(without disparity refinement); (d) mismatched map (without disparity refinement); (e) disparity maps obtained by proposed algorithm (disparity refinement); (f) mismatched map (disparity refinement)
Parameter | ωk | Tg | ε | σ | Th | ω1 | Tcen | λG | ω2 |
---|
Value | 9 | 4.335 | 1×10-5 | 1.5 | 1 | 0.9 | 45 | 0.065025 | 0.1 |
|
Table 1. Parameters involved in proposed stereo matching algorithm
Algorithm | SAD | CT | MCT | MCT’ | GRD | Proposed algorithm |
---|
Tsukuba | 4.65 | 3.54 | 4.58 | 4.50 | 2.72 | 2.53 | Venus | 3.45 | 2.00 | 3.63 | 3.55 | 1.68 | 1.60 | Teddy | 14.21 | 8.56 | 13.05 | 11.32 | 7.45 | 7.57 | Cones | 7.70 | 4.87 | 7.49 | 6.78 | 4.45 | 4.04 | Avg | 7.50 | 4.74 | 7.19 | 6.54 | 4.08 | 3.93 |
|
Table 2. Percentage of false match in no-occluded region of different matching cost algorithms%
Algorithm | SAD | CT | MCT | MCT’ | GRD | Proposed algorithm |
---|
Tsukuba | 4.97 | 3.12 | 3.36 | 3.59 | 3.59 | 2.86 | Venus | 8.34 | 5.25 | 7.17 | 7.15 | 4.12 | 4.13 | Teddy | 25.54 | 19.28 | 23.48 | 22.76 | 17.56 | 17.27 | Cones | 23.40 | 17.04 | 19.71 | 18.88 | 16.13 | 16.06 | Avg | 15.56 | 11.17 | 13.43 | 13.10 | 10.35 | 10.08 |
|
Table 3. Percentage of false match in all regions of different matching cost algorithms%
Algorithm | Tsukuba | Venus | Teddy | Cones | Average |
---|
No-occluded | All | No-occluded | | All | No-occluded | All | No-occluded | All |
---|
BPcompressed[21] | 2.68 | 3.63 | 1.33 | 1.89 | 8.36 | 13.9 | 3.71 | 9.85 | 5.67 | GC-occ[5] | 1.19 | 2.01 | 1.64 | 2.19 | 11.20 | 17.40 | 5.36 | 12.4 | 6.67 | AdaptAggrDP[22] | 1.57 | 3.50 | 1.53 | 2.69 | 6.79 | 14.3 | 5.53 | 13.2 | 6.14 | RTCensus[23] | 5.08 | 6.25 | 1.58 | 2.42 | 7.96 | 13.8 | 4.10 | 9.54 | 6.34 | FastAggreg[24] | 1.16 | 2.11 | 4.03 | 4.75 | 9.04 | 15.2 | 5.37 | 12.6 | 6.78 | SemiGlob[25] | 3.26 | 3.96 | 1.00 | 1.57 | 6.02 | 12.2 | 3.06 | 9.75 | 5.10 | GradAdaptWgt[26] | 2.26 | 2.63 | 6.99 | 1.39 | 8.00 | 13.10 | 2.61 | 7.67 | 5.58 | Proposed algorithm | 2.52 | 3.23 | 0.23 | 1.05 | 5.70 | 12.35 | 3.29 | 9.75 | 4.74 |
|
Table 5. False matching rates of different stereo matching algorithms in no-occlusion region and all regions%
Stereopairs | False matching rate /% |
---|
MCT[16] | MCT'[17] | MST[6] | AW[13] | GF[14] | CT-GF[15] | CT-MST[15] | Proposed algorithm |
---|
AloeArtBaby1Baby2Baby3BooksBowling1Bowling2Cloth1Cloth2Cloth3Cloth4DollsFlowerpotsLampshade1Lampshade2LaundryMidd1Midd2MoebiusMonopolyPlasticReindeerRocks1 | 9.4918.614.296.026.9912.2618.2211.151.985.794.053.5811.4314.2724.7327.6024.7748.7547.7014.1628.4944.4013.835.38 | 9.3119.213.915.986.7211.3417.2311.161.875.214.143.6412.1515.2125.4327.7725.2140.1440.0113.8627.9739.4312.445.19 | 6.6113.747.6514.129.5312.9719.1412.631.185.403.202.537.8617.2711.2914.2019.3918.9620.2111.3316.8530.3011.964.14 | 6.5012.885.4513.648.7612.6516.8710.431.255.563.813.216.4315.4412.3617.4416.8936.4735.5613.2615.4332.4411.735.43 | 7.0612.033.163.974.7110.216.218.181.884.202.792.257.129.6312.1212.4820.6436.9534.0210.6419.7427.067.843.97 | 7.0412.163.314.034.979.966.188.471.964.832.802.387.279.6710.9310.5418.5433.4532.9610.9218.9922.228.154.03 | 6.5713.867.9113.679.4712.4719.5812.581.155.363.162.407.6217.0511.3714.0817.8619.2119.3211.6616.6431.0611.923.97 | 5.9310.162.532.663.918.016.075.881.243.432.291.645.267.5310.0411.1615.5131.7631.909.5719.5025.806.892.86 | Rocks2Wood1Wood2Average | 4.136.494.4115.67 | 3.247.174.3714.79 | 3.3311.174.2011.52 | 5.028.996.1512.59 | 2.674.932.7110.34 | 2.744.932.959.87 | 3.9810.984.1211.45 | 1.943.722.188.67 |
|
Table 6. False matching rate of different stereo matching algorithms in no-occluded region
Method | MCT[16] | MCT'[17] | MST[6] | AW[13] | GF[14] | CT-GF[15] | CT-MST[15] | Proposed |
---|
Averagerunning time /s | 3.97 | 4.35 | 1.56 | 15.87 | 4.76 | 6.75 | 4.12 | 6.63 |
|
Table 7. Average running time of 31 Middlebury stereo pairs with different algorithmss