Analysis of the proficiency of fully connected neural networks in the process of classifying digital images. Benchmark of different classification algorithms on high-level image features from convolutional layers

Highlights：

• Research that investigates CNN architectures for classification task.

• Two steps method that combines CNN with other ML classifiers.

• Large experimental phase considering different architectures and datasets.

• Results used to provide guidelines on the choice of the best classifiers.

• CNNs can improve their performance by using the proposed method.