Machine Learning with Images
Jason Bell · Machine Learning · 2020
This chapter looks at image processing and classification, starting with using a basic neural network and then extending that knowledge to use convolutional neural networks for image classification. It first uses the multilayer perceptron to work on the Modified National Institute of Standards and Technology number data set and then introduces the more complete convolutional neural network for a more thorough way of extracting features from image data. In its basic form, a computer-based image is a grid of numbers. Each “square” is called a pixel. The chapter presents an example of an 8 pixel by 8–pixel image. It shows the image information that can be handled depending on the image depth; the larger the depth, the more colors that can be introduced. In the context of machine learning, it may be prudent to reduce the color depth to speed up training; reducing the image size will help too.