Human Action Recognition Using Joint Positions from Depth Videos
Adarsh S1, Asha S2
1Adarsh S, PG Scholar, Department of Electronics and Communication, SCT College of Engineering, Trivandrum (Kerala), India.
2Asha S, Asst. Prof., Department of Electronics and Communication, SCT College of Engineering, Trivandrum (Kerala), India.
Manuscript received on 15 June 2015 | Revised Manuscript received on 25 June 2015 | Manuscript Published on 30 June 2015 | PP: 169-173 | Volume-4 Issue-5, June 2015 | Retrieval Number: E4122064515/15©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Human Action Recognition using visual information in a given image or sequence of images, has been an active area of research in computer vision applications. The image captured by conventional camera does not provide the suitable information to perform comprehensive analysis. However, depth sensors have recently made a new type of data available. Most of the existing work focuses on body part detection and pose estimation. A growing research area addresses the recognition of human actions based on depth images. In this paper, the following contributions are made: the proposed method makes an efficient representation of human actions by constructing a feature vector based on the human’s 3D joint positions. These locations are extracted fromdepth videos which are taken with the help of Microsoft Kinect sensor. Experiments were performed on a new dataset Kinect Action Dataset (KAD-10). The data set consists of 3D sequences of 10 indoor activities performed by 10 individuals in varied views. Then these feature vectors are given to K-Nearest Neighbour (KNN) classifier to perform the action classification task which results in action labels.
Keywords: Video Surveillance, Depth Sensor, Body Part Labeling, Depth Image Features, Randomized Decision Forest, Joint Position Estimation, K-Nearest Neighbour Algorithm.
Scope of the Article: Image analysis and Processing