Table 1
Results on testing different model augmentation and input view sizes.
| Model | Avg. precision |
|---|---|
| 201/2001 bins | 92.0 ± 0.7% |
| 101/1001 bins | 92.7 ± 0.7% |
| Without white noise | 89.6 ± 0.7% |
| Without phase-inversion | 90.4 ± 0.7% |
| Without phase-shifts | 90.5 ± 0.7% |
| Without any augmentation | 85.2 ± 0.7% |
Notes.The 101-bin and 201-bin models are with all three methods of data augmentation. Testing of individual augmentation techniques was performed by removing each individual method in turn from the 101-bin model. Four-fold cross validation was used for this testing, and the numbers given are on the validation dataset.
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.