Robust Feature Based Automated Multi-View Human Action Recognition System Using Machine Learning

Authors

  • M. Mahalakshmi
  • M. Dayanidhy

Abstract

An automatic human action recognition technique is proposed in this paper. The main intention of this paperwork is to provide a new approach for image recognition using an artificial neural network. In multimedia processing and traditional techniques, human-machine interaction is a research topic for communication that allows disabled people to communicate easily with machines. Human body shapes which cause the change in appearance is also in the study for human action recognition. Our work focuses around human action recognition from a video scene. The given input image is pre-processed by Gaussian filter and a mean Eigen space is produced by taking a mean of the similar postures. In addition to these observances, we used Artificial Neural Network which produces the robustness.

Downloads

Published

2020-01-19

Issue

Section

Articles