Orthogonal Variant Moments in Computer Vision

The Orthogonal Variant Moments (OVM) are proposed in this paper as a way of characterization of any function or signal in general. This approach provides many benefits over the traditional characterization by invariant moments when it is applied to images, time series, wave recognition and information extraction, as it provides relevant information. Our approach to the theory of visual perception is based on the study of the low level vision system by OVM while most of the works on this field use invariant moments. The OVM is shown to be a powerful tool to extract general information of any signal including ND-signals. An application of this method to computer vision proves the efficiency of this approach.