Real-time and Simultaneous Recognition of Multiple Moving Objects Using Cubic Higher-order Local Auto-Correlation

Abstract

Real-time recognition of moving objects is an important problem in video surveillance applications, ITS (Intelligent Transport Systems), robot vision and so on. In this paper, we propose a method to recognize multiple moving objects simultaneously by using Cubic Higher-order Local Auto- Correlation (CHLAC) features. To perform the recognition, we exploit the additivity property of CHLAC which allows us to express the feature values as a linear coupling between the features of each object. The effectiveness of the method is verified by performing recognition and counting the number of pedestrians. We also show that our method can be robust to changes in object scale and speed.