Human Computer Interaction Method for Feeding Demands of the Disabled and the Sick Based on Micro Facial Posture Capture

There are a large number of elderly, disabled, and bedridden people who can not eat independently in China. With the development of science and technology, many eating auxiliary robot systems have been studied at home and all over the world. However, most of these systems lacks humanized humancomputer interaction relying only on based buttons, joysticks or human voice. Such methods can not be well applied to groups with limited limbs, weak body and tired mind. In this paper, we propose a food demand recognition method for the disabled and the sick based on micro facial posture capture. This method can capture the key points of face and face in real-time and accurately, so as to identify whether the user is eating, ready for new food or refusing to eat the food picked up by the robot, and optimize the human-computer interaction method. The effectiveness of the proposed method is proved by experiments.