Multi-class Classification on Riemannian Manifolds for Video Surveillance

In video surveillance, classification of visual data can be very hard, due to the scarce resolution and the noise characterizing the sensors’ data. In this paper, we propose a novel feature, the ARray of COvariances (ARCO), and a multi-class classification framework operating on Riemannian manifolds. ARCO is composed by a structure of covariance matrices of image features, able to extract information from data at prohibitive low resolutions. The proposed classification framework consists in instantiating a new multi-class boosting method, working on the manifold Sym+d of symmetric positive definite d × d (covariance) matrices. As practical applications, we consider different surveillance tasks, such as head pose classification and pedestrian detection, providing novel state-of-the-art performances on standard datasets.