Algorithm for Predicting Pedestrian Behavior on Public Roads
Marina Moseva, Ksenia Polyantseva, Mikhail Gorodnichev, Artem Pavlikov
Abstract
Human Pose Estimation (HPE) is a way to identify and to classify the nodes of the human body. Essentially, it is a way of determining coordinates of each node (arm, head, torso, etc.) called a key point and defining the position of the human body. The relationship between these points is called a pair. The article analyzes existing approaches to predicting human behavior and compares them. The article contains a comparison and description of tools and technologies for the implementation of human behavior detection systems. It describes the implementation of the system, the development of a module to detect such a sign of deviant behavior as falling, the development of an application that acts as a client application to display notifications of the detection of human falls, manual testing of the implemented program.