Pixel-based object detection and tracking with ensemble of support vector machines and extended structural tensor

Abstract

In this paper we propose a system for visual object detection and tracking based on the extended structural tensor and the ensemble of one-class support vector machines. First, the input color image is transformed with the anisotropic process into the extended structural tensor. Then the tensor space is clustered into the number of partitions which are used to train a corresponding number of one-class support vector machines composing an ensemble of classifiers. In run-time the ensemble classifies the input video stream into an object and background. Thanks to high discriminative properties of the extended structural tensor and to the diversity of the ensemble of classifiers the method shows very good properties which were shown by experiments on real video sequences. © 2012 Springer-Verlag.