Sensor based human action recognition and known public datasets a comprehensive survey

Abstract

Human activity recognition (HAR) using specific information collected from many data acquisition devices such as (camera in video-based activity recognition or sensor in sensor-based activity recognition) which is employed in many types of research domains such as human monitoring, healthcare, and computer-human interaction. This paper provides an overview of the latest papers sensor-based activity recognition and known public datasets used in such papers. Which will gives a diagram of the most recent parts of the accompanying perspectives: significant sign, information catch, and preparing, techniques for managing obscure areas and patterns on the body, choice of fitting highlights, movement models, exercise manuals and measures to evaluate action execution and strategies for surveying the ease of use of the HAR framework. The study covers the identification of repetitive exercises, cases, falls and dormancy.