HUMAN ACTION RECOGNITION IN A VIDEO BASED ON SPATIO-TEMPORAL FEATURES |
Author(s): |
| Sushmitha.V.C , Dayananda Sagar college of engineering; Rashmi.S.R, Dayananda Sagar college of engineering |
Keywords: |
| Human Action Recognition (HAR), STIP, KNN Classifier |
Abstract |
|
Recognizing human actions in complex scenes is a challenging problem due to background clutters, camera motion, occlusions, and illumination variations. There are several realistic scenarios where human action recognition (HAR) highlights its importance for security purpose. Challenges in HAR are Application Domain, Variations in Inter and Intra class, Background and Recording settings, Human variation, Action variation…etc. In this paper spatio temporal interest points (STIP) are located in region that shows a high variation of image intensity in all three direction (x,y,t). Spatio temporal corners are located at spatio corners such that they invert motion in two consecutive frame. To classify the frames in video we are using K-nearest neighbor (KNN) by loading the trained features. Experimental results shows the effectiveness of our model. |
Other Details |
|
Paper ID: IJSRDV5I31132 Published in: Volume : 5, Issue : 3 Publication Date: 01/06/2017 Page(s): 1413-1416 |
Article Preview |
|
|
|
|
