Most of the state-ofarts visual object classification methods use bag of words model for image representation. In this method, patches extracted from images are described by different shape and texture descriptors such ...