Regularized Deep Convolutional Neural Networks for Feature Extraction and Classification

Deep Convolutional Neural Networks (DCNNs) are the state-of-the-art in fields such as visual object recognition, handwriting and speech recognition. The DCNNs include a large number of layers, a huge number of units, and connections. Therefore, with the huge number of parameters, overfitting can occur. In order to prevent the network against this problem, regularization techniques have been applied in different positions. In this paper, we show that with the right combination of applied regularization techniques such as fully connected dropout, max pooling dropout, L2 regularization and He initialization, it is possible to achieve good results in object recognition with small networks and without data augmentation.