Human-Machine Interaction based on Eye Command: Algorithm to detect the movement of the eyes and recognize the command

Abstract

Several features on the face can be identified, like eyes, nose, mouth, lips, etc. These features can be used for many purposes. The most important of which is helping people with disabilities who lose one or all of their senses. This paper presents an algorithm that enables determining the direction of the person’s eye in order to manage an intelligent interface where it determines the desired direction. So, that detects the face with the camera, and through the detector MTCNN (Multi-Task Cascaded Convolutional Neural Networks), can be identified the features, and with a map (called landmark points), it is possible to know the coordinates of every feature in the face. We need to define the center(pupil) of the eye, which with his coordinates makes the decision, and the pupil’s coordinates calculate through the map and a set of calculations