Going Deeper into Action Recognition: A Survey

Samitha Herath, Mehrtash Harandi, Fatih Porikli

Published 2016-05-16Version 1

We provide a detailed review of the work on human action recognition over the past decade. We refer to "actions" as meaningful human motions. Starting with methods that are based on handcrafted representations, we review the impact of revamped deep neural networks on action recognition. We follow a systematic taxonomy of action recognition approaches to present a coherent discussion over their improvements and fall-backs.