Fig. 5.

Accuracy versus epoch graphs evaluated using multi-scale deep CNN for training and validation data of the best results obtained from the 6 models: (a) multi-class classification of dataset A with 5-fold CV using mode 3 and mode 4. (b) multi-class classification of dataset B with 5-folds using mode 4, mode 5 and mode 6. (c) multi-class classification of dataset B with hold-out validation using mode 4, mode 5 and mode 6. (d) binary classification of dataset A with 5-fold CV using mode 1 and mode 2. (e) binary classification of dataset B with 5-fold CV using all seven modes. (f) binary classification of dataset B with hold-out validation using the first six modes.