Classification of objects in residential monitoring systems

This paper presents a real-time object classification method for distinguishing between humans, pets and other objects in residential security systems. Specifically, we propose using an ensemble of radial basis function (RBF) networks on gradient images extracted from the scene. A specific advantage of using an ensemble is its ability to cope with the inherent variability in the image formation and data acquisition. process. Experimental results for two different ensembles of networks are presented with average cross validation performances of 91% and 95% respectively