IMAGE STITCHING BASED ON PARAMETRIC TRANSFORMATIONS

Abstract

Transformation functions play a major role in mosaic images of different geometry. In this article, parametric global transformations, such as the similarity model, affine model and perspective projection model (projective model) are explored, and their performance in the mosaic of images with geometric differences are compared with respect to the execution time and registration quality. The approaches under study are tested on different types of images to deduce a scientific interpretation of results.