Determining the head position regarding the digital video camera installed on a computer

Authors: Kataev M. Yu., Mosyaev A. E.

Annotation: Currently, applications that require tracking the position of the human head (for example, video conferences, computer games, virtual reality, etc.) are gaining more and more popularity. Especially stimulated are studies aimed at solving computer vision problems for finding head movements in real time. A feature of the proposed solution is that head tracking can be performed without the use of expensive video cameras that realize stereo or more complex measurements, but with just one digital camera. In this article, the authors propose a technique based on the computer vision approach that allows detecting detect head movements of a user sitting in front of computer or laptop. The technique relies on a person facemask, obtained using a pretrained convolutional neural network. The resulting mask is converted to the required format, relative to which, further, the key points of the face used to evaluate the inclinations and turns of the head. The article describes the structure of the developed program and the results of measurement processing.

Keywords: human head, computer vision, image, processing methods