Open AccessA Review on Classification Techniques for Human Activity Recognition
Author(s) -
Sonali Bhardwaj,
Ashok Kumar Bathla
Publication year - 2015
Publication title -
journal of advance research in business, management and accounting
Language(s) - English
Resource type - Journals
ISSN - 2456-3544
DOI - 10.53555/nnbma.v1i2.131
Subject(s) - computer science , viewpoints , artificial intelligence , frame (networking) , support vector machine , task (project management) , activity recognition , entertainment , machine learning , computer vision , pattern recognition (psychology) , bag of words model , engineering , art , telecommunications , systems engineering , visual arts
Recognizing human actions from video sequences has many important applications like video surveillance, patient monitoring, human computer interaction, dance choreography analysis, analysis of sports events and entertainment environments. It involves processing the video into frames firstly and finding out the interest points, then extracting the features and lastly specifying and labeling the videos following an appropriate classifying approach like Support Vector Machine, bag of words or nearest neighbor. This paper provides a detailed overview of various state-of-the-art research papers on human activity recognition using different types of classifiers. We surveyed various challenges exhibited by computer vision researchers like the problem of occlusion, 2D/3D pose estimation, variations in viewpoints, human body modeling especially of a person who is paralyzed or injured. From this survey, we can make conclusion of various advantageous and disadvantageous facts about different classifiers used in the detection and classification task.