A Comprehensive Review of Deep Learning for Activity Recognition

Human Activity Recognition (HAR) is an important field in smart healthcare that has attracted remarkable attention from researchers. Several HAR systems are existing in the literature. However, there exist substantial challenges that could influence the performance of the recognition system in practical scenarios. Recently, as deep learning has demonstrated its effectiveness in many areas, plenty of deep methods have been investigated to address the challenges in activity recognition. This survey aims to provide a comprehensive overview of HAR approaches based on deep learning. Our work discusses the challenges of HAR systems to provide a comprehensive overview for researchers who are interested in this field of HAR. Firstly, we identify the challenges of HAR. Then, we discuss various sensors to implement the HAR systems. Finally, we discuss the challenges and role of deep learning in HAR. We also compare the performance of recently proposed methods on popular benchmark datasets. We review 22 benchmark datasets for human action recognition. Some potential research directions are discussed to conclude this survey.