Python Image Processing Cookbook

How it works...

If you choose a random face from the test dataset and predict the class (here, it is the name of the mathematician) along with the probability of prediction, you will get an output like this:

As you can see, the face of Bayes was recognized correctly with a probability of around 80% using the linear SVM classification model trained on the face embeddings.

The face identification problem can be converted into a multi-class classification problem in the embedding space by using the following steps:

1. First, extract the high-quality face-embedding features from a set of labeled training images of the mathematicians with the FaceNet ...