Human activity recognition (HAR): techniques, applications, and future directions
摘要
Human Activity Recognition (HAR) involves the development of intelligent methods to classify and recognize human actions and behaviour in real-world environments by analyzing data acquired from various sensors and sources, extracting meaningful patterns, and making informed decisions about the identified activities. This discipline focuses on training models to understand and interpret human movements using specialized sensors and cameras capable of detecting actions such as walking, sitting, or gesturing. The ultimate aim of HAR is to enhance technology’s ability to comprehend human behaviour, contributing to the development of smarter, more responsive devices for domestic applications and ensuring greater safety. This article provides a concise overview of the significance and purpose of HAR, highlighting its applications across diverse domains such as healthcare, sports, and surveillance. The primary objective of this paper is to present a comprehensive review of HAR systems, offering valuable insights to guide new researchers, inform future studies, and deepen the overall understanding of this evolving field.