Table 1.
Comparison of the 3D reconstruction results for the pen.
| Scale 1 | Scale 2 | ||||||||||||||
| Orientations | 0 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 2 | 3 | 4 | 5 | 6 | 7 | 8 |
| Nev | 100 | 71 | 65 | 78 | 77 | 87 | 100 | 105 | 73 | 78 | 85 | 98 | 121 | 128 | 146 |
| Nm | 28 | 15 | 14 | 15 | 14 | 16 | 17 | 18 | 15 | 15 | 16 | 18 | 22 | 24 | 27 |
| Matching rate | 28 | 21 | 21 | 19 | 19 | 18 | 17 | 17 | 21 | 20 | 19 | 18 | 18 | 19 | 19 |
| Isolated events | 2.9 | 5.6 | 6.4 | 5.4 | 5.8 | 5.1 | 4.5 | 4.1 | 5.1 | 4.9 | 4.7 | 4.1 | 3.0 | 2.6 | 2.1 |
| Merr | 8.0 | 4.1 | 3.9 | 4.2 | 3.9 | 4.1 | 3.9 | 4.1 | 3.6 | 3.6 | 3.7 | 3.6 | 3.6 | 3.6 | 3.6 |
| Nm-correct | 24.9 | 14 | 13 | 14 | 13 | 15 | 16 | 17 | 14 | 14 | 15 | 17 | 21 | 23 | 25 |
| Scale 3 | Scale 4 | ||||||||||||||
| Orientations | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | |
| Nev | 74 | 80 | 87 | 106 | 131 | 154 | 169 | 77 | 79 | 85 | 99 | 129 | 145 | 170 | |
| Nm | 16 | 17 | 17 | 21 | 26 | 31 | 34 | 19 | 19 | 19 | 22 | 30 | 34 | 39 | |
| Matching rate | 22 | 21 | 20 | 19 | 20 | 20 | 20 | 24 | 24 | 23 | 23 | 23 | 23 | 23 | |
| Isolated events | 5.0 | 4.7 | 4.5 | 3.3 | 2.4 | 1.8 | 1.5 | 3.3 | 3.3 | 3.2 | 2.6 | 1.6 | 1.4 | 1.0 | |
| Merr | 5.2 | 5.2 | 5.2 | 5.1 | 4.9 | 4.9 | 5.0 | 8.3 | 8.3 | 8.3 | 8.3 | 8.2 | 8.1 | 7.8 | |
| Nm-correct | 14 | 15 | 15 | 19 | 24 | 29 | 32 | 17 | 17 | 17 | 20 | 27 | 31 | 36 | |
The first column (0 orientations) presents the results obtained applying the matching algorithm to the retinas events (algorithm A, section Event Matching), while the rest of the columns are related to the pair-wise application of the matching algorithm to the outputs of the Gabor filters (algorithm B, section Event Matching), from Scale 1 (smaller kernels) to Scale 4 (larger kernels). For each scale, different numbers of orientations are considered (from 2 to 8), as indicated in the first row (Orientations). Second row (Nev) shows the number of events processed (in Kevents) by the matching algorithm in each case (i.e., the total number of events generated by all the filters). Third row (Nm) presents the number of matched events (in Kevents) produced by the algorithm, while fourth row (Matching Rate) shows the ratio of matched events over the total number of events generated by the Gabor filters (Matching Rate = 100 · Nm/Nev, in %). Fifth row (Isolated events) shows the ratio of isolated events over the total number of matched events (in %). Sixth row (Merr) presents the ratio of wrongly matched events over the total number of matched events (in %). The last row (Nm-correct) encapsulates the number of matched events with the ratio of isolated and wrongly matched events, presenting the number of correctly matched events (Nm-correct = Nm − , in Kevents).