Summary

In this chapter, we looked at how to develop the face detection application using the face_recognition library, which uses the HOG-based model to identify the faces in the images. We have also used the pre-trained convolutional neural network, which identifies the faces from a given image. We developed real-time face recognition to detect the names of people. For face recognition, we used a pre-trained model and already available libraries. In the second part of the chapter, we developed the face emotion recognition application, which can detect seven major emotions a human face can carry. We used TensorFlow, OpenCV, TFLearn, and Keras in order to build the face emotion recognition model. This model has fairly good accuracy for predicting the face emotion. We achieved the best possible accuracy of 67%.

Currently, the computer vision domain is moving quickly in terms of research. You can explore many fresh and cool concepts, such as deepfakes and 3D human pose estimation (machine vision...