A fine grainedresearch over human action recognition

Abstract

Human Action Recognition from videos has been an active research is in the computer vision due to its significant applicability in various real-time applications like video retrieval, human-robot interactions, and visual surveillance, etc. Though there are so many surveys over Human action Recognition, they are limited to various constraints like only focusing on the methods in few orientations only. Unlike the earlier ones, this paper provides a detailed survey according to the basic working methodology of Human action recognition system. Initially, a detailed illustration is given about various standard benchmark datasets. Further, following the methodology, the survey is accomplished in two phases, i.e., the survey over feature extraction approaches and the survey over action classification approaches. Further, a fine-grained survey is also accomplished under every phase based on the individual strategies.