Combined local global optical flow approach for image sequences

Abstract

This paper addresses the optical flow analysis and the smoothing effect in Local and Global Differentiation. There are lots of Local and Global methods and their variants available. But in this paper, as a prototype of local methods least-square fit of Lucas and Kanade has been employed, while the Horn and Schunck approach has been considered as the representative for a global method. Each of these methods has their advantages and shortcomings. Which are mentioned in their respective section. As a consequence, the proposed made from the hybridization algorithm incor-porates the advantages of both field and posses less disadvantages. Improvement is made in our algorithm using the spatio-temporal approach and appropriate differentiation approximation.