Optimized Cascade of Classifiers for People Detection Using Covariance Features

People detection on static images and video sequences is a critical task in many computer vision applications, like image retrieval and video surveillance. It is also one o f most challenging task due to the large number of possible situations, including variations in people appea rance and poses. The proposed approach optimizes an existing approach based on classification on Riemannian man ifolds using covariance matrices in a boosting scheme, making training and detection faster while maintaining equivalent performances. This optimisation is achieved by clustering negative samples before training, providing a smaller number of cascade levels and less weak classifiers in most levels in comparison with the or iginal approach. Our work was evaluated and validated on INRIA Person dataset.