Automatic Activity Recognition for Video Surveillance

Abstract

Activity recognition is having a wide range of applications in automated surveillance and is an active research topic among computer vision community.In this paper, an activity recognition approach is proposed.Motion information is extracted from the difference image based on Region of Interest (ROI) using 18-Dimensional features called Block Intensity Vector (BIV).The experiments are carried out on the KTH dataset considering four activities viz., (walking, running, waving and boxing) with SVM.The approach shows an overall performance of 94.58% in recognizing the actions performed.Experimental results show that the proposed approach is comparable with the existing methods.