Facial Paralysis is one of the neural disorders that is increasing nowadays. Anyone can be infected by this when the cranial nerve is affected. As a result, they will not be able to contract the muscles on the face on either of the sides. This directly affects the social life of patients who has this palsy, as they couldn’t show much expression and communicate naturally like other healthy persons. Computer vision-based algorithms that are based on the CNN model and auto encoders can be used for accurate identification with precision and allow the extraction of high-level information from these facial key points, so that detection of facial paralysis can be used for the betterment of medical diagnosis. In this paper, the MediaPipe framework is used to identify the necessary key points of the facial parts of images containing palsy. The framework provides the key points for entire face which are used in many applications like a face mesh, among those the essential points are the facial parts which are required for analysis of palsy are considered.

Facial key points, contours, palsy, MediaPipe.