Author Affiliations
1School of Electronic and Information Engineering, Liaoning Technical University, Huludao, Liaoning 125105, China2State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China3Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang, Liaoning 110016, Chinashow less
Fig. 1. Direct relevant neighborhood of matching points
Fig. 2. Window shape and weight allocation
Fig. 3. Main flow chart of matching in this paper
Fig. 4. Sample datasets diagrams (Jadeplant). (a) Left image; (b) right image; (c) ground truth; (d) mask
Fig. 5. Mismatching rate of different window shapes
Fig. 6. Different types of matching windows
Fig. 7. Example of the effect improvement of proposed algorithm and the original method. (a) Left images; (b) ground truth images; (c) ELAS; (d) proposed algorithm
Fig. 8. bad index comparison
Fig. 9. avgErr index comparison chart
Fig. 10. Sample graphs of effectiveness improvement. (a) Left images; (b) ground truth; (c) GA-Net; (d) GwcNet; (e) Census; (f) SGM; (g) proposed algorithm
Fig. 11. Self-collecting images. (a) Left images; (b) GA-Net; (c) GwcNet; (d) Census; (e) SGM; (f) proposed algorithm
Data name | Ebad /% | Eavg /% | Data name |
---|
ELAS | | Proposed | ELAS | Proposed |
---|
Adirondack | 12.070 | 11.890 | 0.880 | 0.877 | Adirondack | ArtL | 19.800 | 19.270 | 1.760 | 1.630 | ArtL | Jadeplant | 34.680 | 34.660 | 1.170 | 1.170 | MotorcycleE | MotorcycleE | 13.450 | 13.210 | 2.210 | 2.080 | PianoL | PianoL | 37.520 | 36.530 | 1.960 | 1.930 | Pipes | Pipes | 15.300 | 14.760 | 1.900 | 1.800 | layroom | Playroom | 26.680 | 26.040 | 4.550 | 4.270 | Playtable | Playtable | 26.380 | 26.290 | 0.930 | 0.920 | PlaytableP | PlaytableP | 14.840 | 14.740 | 0.890 | 0.890 | Recycle | Recycle | 17.850 | 16.850 | 0.910 | 0.910 | Teddy | Vintage | 33.690 | 33.660 | 2.430 | 2.430 | Vintage |
|
Table 1. Performance comparison of Gaussian weight window under two indices
Data name | Ebad /% | Eavg /% | Data name |
---|
ELAS | | Proposed | ELAS | Proposed |
---|
ArtL | 19.80 | 19.73 | 0.88 | 0.86 | Adirondack | Jadeplant | 34.68 | 29.19 | 8.85 | 6.33 | Jadeplant | MotorcycleE | 13.45 | 13.09 | 1.17 | 1.13 | MotorcycleE | Piano | 22.04 | 21.92 | 2.21 | 2.22 | PianoL | PianoL | 37.52 | 37.50 | 1.96 | 1.97 | Pipes | Pipes | 15.30 | 15.19 | 0.93 | 0.92 | PlaytableP | Playroom | 26.68 | 25.63 | 0.91 | 0.89 | Teddy | PlaytableP | 14.84 | 14.79 | - | - | - | Recycle | 17.85 | 17.24 | - | - | - |
|
Table 2. Performance comparison of Scharr filter under two indices
Data name | Ebad /% | Eavg /% | Data name |
---|
ELAS | | Proposed | ELAS | Proposed |
---|
Adirondack | 12.07 | 11.81 | 0.88 | 0.83 | Adirondack | ArtL | 19.80 | 19.55 | 1.76 | 1.74 | ArtL | Jadeplant | 34.68 | 27.85 | 8.95 | 5.99 | Jadeplant | MotorcycleE | 13.45 | 13.30 | 1.17 | 1.15 | Motorcycle | Piano | 22.04 | 21.85 | 1.17 | 1.16 | MotorcycleE | PianoL | 37.52 | 36.03 | 2.21 | 1.94 | PianoL | Pipes | 15.30 | 15.30 | 1.96 | 1.94 | Pipes | Playroom | 26.68 | 25.63 | 0.93 | 0.92 | PlaytableP | Recycle | 17.85 | 17.12 | 1.97 | 1.92 | Shelves | Teddy | 11.42 | 10.88 | 0.91 | 0.83 | Teddy | Vintage | 33.69 | 32.33 | 2.43 | 2.39 | Vintage |
|
Table 3. Performance comparison of proposed algorithm under two indices