Action Recognition by Local Space-Time Features and Least Square Twin SVM (LS-TSVM)

In this research a new approach for human action recognition is proposed. At first, local space-time features extracted which recently becomes a popular video representation. Feature extraction is done with use of Harris detector algorithm and Histogram of Optical Flow (HOF) descriptor. Then we appl...

Full description

Saved in:
Bibliographic Details
Main Authors: Mozafari, K., Nasiri, J. A., Charkari, N. M., Jalili, S.
Format: Conference Proceeding
Language:English
Subjects:
Accuracy
Action Recognition
Computer vision
Feature extraction
Harris
Histogram of Optical Floow (HOF)
Histograms
Humans
KTH dataset
LS-TSVM
Support vector machines
Training data
Twin Support Vector Machines
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this research a new approach for human action recognition is proposed. At first, local space-time features extracted which recently becomes a popular video representation. Feature extraction is done with use of Harris detector algorithm and Histogram of Optical Flow (HOF) descriptor. Then we apply a new extended SVM classifier called least square Twin SVM (LS-TSVM). LS-TSVM is a binary classifier that does classification by use of two nonparallel hyperplanes and it is four times faster than the classical SVM while the precision is better. We investigate the performance of LS-TSVM method on a total of 25 persons on KTH dataset. Our experiments on the standard KTH action dataset shown that our method improves state-of-the-art results by achieving 95.8%, 96.3% and 97.2%% accuracy in case of 1-fold , 5-fold and 10-fold cross validation.
DOI:10.1109/ICI.2011.55