Detection Of Abnormal Visual Events Using HOFO And KNN

The aim of this paper is to detect abnormal events in video streams, a challenging but important subject in video surveillance. A novel algorithm is proposed to address this problem. The algorithm is based on an image descriptor and a nonlinear classification method. The images are subjected to Otsu’s method for global thresholding. A histogram of optical flow orientation as a descriptor encoding the moving information of each video frame is used here. The k-nearest neighbor (kNN) classification algorithm, following a learning period characterizing the normal behavior of training frames, detects abnormal events in the current frame. Further, a fast version of the detection algorithm is designed by fusing the optical flow computation with a background subtraction step. Finally a method to detect abnormal events on several benchmark data set is applied. Detection Of Abnormal Visual Events Using HOFO And KNN Paper ID IJIFR/ V2/ E9/ 045 Page No. 3196-3210 Subject Area Computer Science Engineering