Affine estimation via region expansion

In this work, we present a novel method for accurate affine transformation estimation of image regions. We illustrate the benefits of using such a method in a point matching mechanism that enables locating large amount of point matches with high geometric precision and low rate of false matches. Recent publications have shown that considering the affine transformation model of local regions, is extremely beneficial for the purpose of point matching. Yet, those methods are not used extensively for two reasons. First, because they are computationally more demanding, second reason is that the estimation of the Affine transformations are of limited accuracy and therefore the usage of them is limited. We propose a region expansion method, based on accurate estimation of the affine transformation, that is proven to accurately predict locations beyond the initial local regions. We also prove that expansion beyond local regions is crucial for getting much more accurate transformation estimations. Finally, by reducing the amount of false matches considerably, we reduce the need for computationally demanding post processes.