Background
Given a video sequence, the proposed approach represents the video as a set of cuboids, those cuboids define an event. The video is broken into a set of events, each represented by a group of spatiotemporal cuboids. Specifically, the video is represented as with each event Xi composed of a group of cuboids, i.e.,     where  is the total number of cuboids within the frame.

Motivation
This unusual event detection algorithm adopts a representation based on spatio-temporal cuboids, to detect salient points within the video and describe the local spatio-temporal patch around the detected interest points, and describes each detected interest point with histogram of gradient (HoG) and histogram of optical flow (HoF).