Summary

In this chapter, we started off by seeing why the face is considered an important source that's used to identify different emotions, features, and so on. We also discussed why we cannot have a simple face-matching algorithm because of the large variety of faces available to us. We then approached the topic of Haar Cascades and built several applications using one or multiple cascade classifiers. We also talked in detail about the importance of experimentation when it comes to obtaining a set of arguments that will give us the best results using cascade classifiers. After that, we discussed the GrabCut technique and built a human body segmentation application using a rectangular ROI mask. After that, we modified the code to allow the user to correct the mask to obtain better results. We then covered how to carry out skin segmentation using GrabCut and finally ended the chapter with the emoji filter, which can be run on an input video taken from a webcam or a video file...