Predicting facial key points using a deep neural network

In this section, we will discuss the end-to-end pipeline for facial key-point detection. Facial key-point detection is a challenge for computer vision as it requires the system to detect the face and obtain meaningful key point data, plotting this data on the face and developing a neural network to predict the key-points on the face. This is a difficult problem compared to object detection or image classification, as it first requires facial detection within a bounding box and then key-point detection. Normal object detection just involves the detection of four points representing the four corners of the rectangular bounding box around the object, but key-point detection requires multiple points (more than 10) at various orientations. Extensive key-point detection data and a tutorial on its use can be found at https://www...