Summary

In this chapter, we learned how an Automatic License Plate Recognition program works, and its two important steps: plate localization and plate recognition.

In the first step we learned how to segment an image looking for patches where we can have a plate, and how to use a simple heuristics and Support Vector Machine algorithm to make a binary classification for patches with plates and no plates.

In the second step we learned how to segment with the Find Contours algorithm, extract feature vector from each character, and use an Artificial Neural Network to classify each feature in a character class.

We also learned how to evaluate a machine algorithm with training using random samples and evaluate it using different parameters and features.