Learning Data Mining with Python - Second Edition

Drawing basic CAPTCHAs

Before we can start classifying CAPTCHAs, we first need a dataset to learn from. In this section, we will be generating our own data to perform the data mining on.

In more real-world applications, you'll be wanting to use an existing CAPTCHA service to generate the data, but for our purposes in this chapter, our own data will be sufficient. One of the issues that can arise is that we code in our assumptions around how the data works when we create the dataset ourselves, and then carry those same assumptions over to our data mining training.

Our goal here is to draw an image with a word on it, along with a shear transform. We are going to use the PIL library to draw our CAPTCHAs and the scikit-image library to perform ...