Human Action Recognition STIP Using NB, KNN and LSTM, CNN

Abstract

This paper compares three practical, reliable,
and generic systems for multi view video-based human
action recognition, namely, the NB, KNN and LSTM,
CNN. To describe the different actions performed in
different views, view-invariant features are proposed to
address multi view action recognition. These features
are obtained by extracting the holistic features from
different temporal scales which are modelled as points
of interest which represent the global spatial-temporal
distribution. Experiments and cross-data testing are
conducted on the datasets. The experiment results show
that the proposed approach outperforms the existing
methods on the datasets. INDEX TERMS: Multi-view
video, action recognition, feature extraction,
background subtraction, classification, machine
learning.