Human Activity Recognition and Prediction: Overview and Research Gaps

With several applications in the disciplines of healthcare, human-computer interaction, assistive learning, and many others, Human Actions Recognition, or HAR, is a trending research area. Despite the fact that this field has seen a lot of development over the past ten years, there is still a need to develop strong hybrid machine learning models for human actions recognition that must meet the objectives of the application, must have solid prediction and high recognition rate. A little work has been done in anomaly detection while performing an action. This paper presents a detailed literature survey in this area. Future directions are also presented based on the literature which shows that there is still a need of hybrid optimization technique to enhance the model's functionality.