Deep Learning and its Applications using Python

5Convolutional Neural Networks in Tensorflow

Among deep learning models, Convolutional Neural Network is an algorithm, which acts as an artificial neuron (neural network). This neural network [1] has been widely used to deal with image input for image processing/recognition/classification. Initially, it considers an input, provide importance by prioritizing various objects/features using weigh and bias parameters to differentiate among one another. Examples of convolutional neural network are face recognition, image classification, object tracking, natural language processing, speech recognition and so on. In this chapter, the working of convolutional neural network with its architecture and applicability of convolutional neural network using python programing will be discussed.

5.1 CNN Architectures

In CNN architecture, multiple layers are involved is depicted in Figure 5.1.

Convolutional Neural Network [2] consist of important three layers for processing an input. They are i) convolution layer, ii) fully-connected layer, and iii) pooling layer. The number of layers at each level may vary based on the demand of the model being used i.e., for text or image or signal processing. Which means feature extraction, classification, segmentation and for auto correlation of data. Each set of nodes analyze particular part of input to obtain informative data. This model consists of different kind of layers. They are listed below with short description.

1. Convolutional Layer – It consists ...