In this chapter, we learned about the basics of machine learning and applied them to a small sample application. This allowed us to understand the basic techniques that we can use to create our own machine learning application. Machine learning is complex and involves different techniques for each use case (supervised learning, unsupervised, clustering, and so on). We also learned how to create the most typical machine learning application, the supervised learning application, with SVM. The most important concepts in supervised machine learning are as follows: you must have an appropriate number of samples or a dataset, you must accurately choose the features that describe our objects (for more information on image features, go to Chapter 8, Video Surveillance, Background Modeling, and Morphological Operations), and you must choose a model that gives the best predictions.
If we don't get the correct predictions, we have to check each one of these concepts to find the issue.
In the next chapter, we are going to introduce background subtraction methods, which are very useful for video surveillance applications where the background doesn't give us any interesting information and must be discarded so that we can segment the image to detect and analyze the image objects.