Table 2
Method comparison on random positions.
| Method | µ [10−3] | σµ [10−3] | c [10−4] | σc [10¯4] | Simulated area [deg2] | Relative runtime |
|---|---|---|---|---|---|---|
| No cancel | −21.30 | 1.74 | 5.7 | 1.1 | 69.14 | 1.00 |
| Shape local | −21.83 | 0.93 | 7.2 | 0.6 | 2 × 34.57 | 1.00 |
| Both local | −22.59 | 0.65 | 7.0 | 0.4 | 4 × 34.57 | 1.22 |
| Shape global | −23.81 | 0.88 | 7.8 | 0.5 | 2 × 34.57 | 1.00 |
| Both global | −23.54 | 0.54 | 8.0 | 0.3 | 4 × 34.57 | 1.22 |
| Both 0.02 (g) | −27.83 | 2.36 | 6.0 | 0.3 | 4 × 34.57 | 1.22 |
| RM 0.1 (resp.) | −25.02 | 1.10 | 10.5 | 2.1 | 11 × 4.94 | 0.79 |
| RM 0.1 (fit) | −21.91 | 1.06 | 10.1 | 2.1 | 11 × 4.94 | 0.79 |
| RM 0.02 | −27.42 | 1.19 | 5.5 | 0.9 | 2 × 24.69 | 0.71 |
Notes. The relative runtime is given for this specific example where we used 11 200 (= 140 × 20 × 4) scenes for the fit method, 4400 (= 400 × 11) scenes for the response method in the large shear interval, and 4000 (= 2000 × 2) scenes for the response method in the small shear interval. It is always compared to the runtime of no cancellation. In the second last column, we list the simulated area that went into each method. Indicated is always the unique area multiplied by the required additional versions for each method. The RM in the first column denotes the response method. The estimate for Both 0.02 also used global (g) cancellation. If the method is followed by a float number, this denotes the used shear interval. For the differentiation between response (resp.) and fit approach for RM 0.1 see Sect. 5.1.2.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.