In the previous chapter, we talked about the role of a dataset in image classification systems. We saw that any classifier needs to see some examples of the images that it wants to classify so that it can use the domain knowledge to learn its own set of rules. These rules would eventually help the classifier to make predictions for the new images during its course of operation. We also shared a small dataset comprising of 200 images in the previous chapter. Let's now take a closer look at the contents.
As mentioned, the dataset contains 200 images--100 images of male and 100 of female faces. Since our goal here is to recognize the gender of faces, our dataset must have representations from both the gender categories. All the facial images belong to celebrities (politicians, actors, sportspersons, and so on). Here is a list of celebrity names (categorized according to their genders) whose faces appear in the dataset that we have curated for our project:
Male |
Female... |