CNN – Convolutional Neural Networks: A Complete Understanding
Shriram K. Vasudevan, Sini Raj Pulari, Subashri Vasudevan · 2021
This chapter starts with the basics of Overfitting, Underfitting and Appropriate fitting. Then it goes on to explain what is meant by bias and variance. Next it explains the Convolutional Neural Network (CNN) basics and proceeds to explain the convolution operation in detail. In CNN, zero padding is a significant process to maintain the dimension of the input image. The chapter discusses the need, and step by step explanation of doing, zero padding. If zero padding is to maintain dimension, max pooling is a feature that leaves out insignificant pixels in an image. Max pooling is discussed in detail in the chapter. Followed by max pooling, the process to build a CNN model is explained. The chapter ends with key points and a quiz.