Development of a Network Service for Automatic Human Tracking of a PTZ IP Video Camera Using TensorFlow

With the development of online learning resources, many institutions are thinking about the possibility of introducing lecture notes. It is also planned to introduce a survey of students during the delivery of term papers and dissertations for the compilation of an automatic portfolio. To date, such surveys are conducted on stationary cameras or attract mobile video studios. The purpose of this work is to recreate the camera work without human intervention. To solve this problem, existing approaches to tracking people and turnkey solutions for automated research were considered. For work were selected IP security cameras with motion function and TensorFlow library. The paper consists of 59 pages, contains 10 figures, 1 table, 14 listings and 5 applications. Keywords: computer vision; neural networks; automatic human tracking; automated survey; TensorFlow.