Analysis of Objective Functions in the Problem of Estimating Mutual Geometric Deformations of Images~1

摘要

To estimate mutual geometric deformations of images, a comparative analysis is carried out of the efficiency of using the mean square interframe difference (MSID), the interframe correlation coefficient (ICC), and mutual information (MI) as an objective function. A comparative, largely experimental, analysis is carried out of these similarity measures for various brightness distortions of images, and the convergence of estimates of deformation parameters under the application of objective functions for stochastic pseudogradient procedures is investigated.