Abnormal Event Detection at 150 FPS in MATLAB

TLDR

Speedy abnormal event detection is needed to process large volumes of surveillance videos. The study proposes an efficient sparse combination learning framework that exploits inherent redundancy in video structures. The framework reduces the detection problem to a few inexpensive small‑scale least‑squares optimizations, enabling rapid processing. The method attains high detection rates on benchmark datasets at 140–150 FPS on a standard MATLAB desktop without sacrificing accuracy.

Abstract

Speedy abnormal event detection meets the growing demand to process an enormous number of surveillance videos. Based on inherent redundancy of video structures, we propose an efficient sparse combination learning framework. It achieves decent performance in the detection phase without compromising result quality. The short running time is guaranteed because the new method effectively turns the original complicated problem to one in which only a few costless small-scale least square optimization steps are involved. Our method reaches high detection rates on benchmark datasets at a speed of 140-150 frames per second on average when computing on an ordinary desktop PC using MATLAB.