In this chapter, we'll explore edge detection as one of the most fundamental and widely-used techniques in computer vision. Then, we'll look at edge detection in action, using a number of features and images, by building a Java application that detects edges on different images. As a next step, we'll detail how to use edge detection or convolution with colored RGB images so that we can capture even more features from images. We'll present them using several parameters, which will enable us to control the output of the convolution operation. Then, we'll look at a slightly different type of filter, the pooling layers, and one of the most frequently used: the max pooling layer. After that, we'll put all the pieces together for the purpose of building and training a convolution neural network. Finally, we'll use the convolution architecture, as we did in the previous chapter, to optimize handwritten digit recognition with an accuracy of 99.95%.

We will cover the following topics in this chapter:

• Understanding edge detection
• Building a Java edge detection application
• Convolution on RGB images
• Working with convolutional layers' parameters
• Pooling layers
• Building and training a convolutional neural network
• Improving the handwritten digit recognition application