Open Access

Table 4

Hyperparameters of the neural network.

Parameter Value
Architecture
Starting convolution window [7 × 7]
Convolution stride 1
Convolution padding 0
Convolution activation ReLU
Dropout probability 0.3
Pooling type MaxPooling
Pooling size 2 × 2
Number of convolution layers 10
Number of pooling layers 4
Number of dense layers 6
Dense activation function ReLU

Optimization
Batch size 32
Learning rate 2.5 × 10−4
Optimizer Adam
Loss function Categorical crossentropy
Early stopping Δ and patience 10−4, 19 epochs

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.