Human action recognition: An overview

Abstract

In this paper, an approach for detection of human actions from video is given. It has many applications, such as surveillance, human healthcare systems, border security, virtual reality, HMI, etc. This paper includes the detection of human actions and labeling them on an LCD. The frames are made from video and then median filter is applied. The feature extraction is done by applying PCA. The segmentation process is done using Gabor filter. Finally, KNN classifier is applied for classification. Different datasets are also provided.