Fig. 4.

Classifier architecture (see Sect. 3.6 for a detailed description). The neural network is composed of a first convolution layer, a pooling layer and five inception blocks followed by two fully connected layers and a softmax layer. The size and the number of feature maps are indicated for each layer. Further details can be found in Table A.1.
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.