When we said that the solution will follow the basic framework of an image classification system, it was an indirect way of saying that we need to accomplish the following tasks:
Decide upon and extract a suitable set of features from images.
Train a classifier so that it learns a set of rules that discriminate between male and female faces.
We have also discussed the fact that point 2 requires us to have a set of images at hand that we can show to the classifier for the purpose of learning rules. These sets of images are what is known as the dataset. Every image classification task that we wish to accomplish needs a dataset that is tailored to its needs. So, for our gender classification project, we would need a set of images that contain faces of both male and female subjects.
Don't worry, you won't have to download these images. Along with the chapter, we provide you with a set of images that will serve as our dataset. Now that we have the dataset, can we directly...