Synthesis and Optimization of a Stochastic Algorithm for Image Registration Using Shannon's Maximum Mutual Information Criterion

被引：0

作者：

Tashlinskii, A. G. ^{[1
]}

Ibragimov, R. M. ^{[1
]}

Safina, G. L. ^{[2
]}

机构：

[1] Ulyanovsk State Tech Univ UlSTU, Dept Radio Engn Telecommun & Informat Secur, Ulyanovsk 432027, Russia

[2] Moscow State Univ Civil Engn Natl Res Univ, Dept Fundamental Educ, Moscow 129337, Russia

来源：

OPTOELECTRONICS INSTRUMENTATION AND DATA PROCESSING | 2024年 / 60卷 / 02期

基金：

俄罗斯科学基金会;

关键词：

image; estimation; parameter; algorithm; image registration; stochastic procedure; convergence; Shannon mutual information; SAR;

D O I：

10.3103/S8756699024700213

中图分类号：

O4 [物理学];

学科分类号：

0702 ;

摘要：

An algorithm for estimating image registration parameters using Shannon's maximum mutual information criterion is synthesized on the basis of the mathematical methods of stochastic gradient adaptation. Expressions for the gradient of mutual information are obtained. At the same time, the derivatives of the entropy of images with respect to the estimated parameters are found using the Parzen window method. The synthesized algorithm is optimized to improve performance.

引用

页码：181 / 189

页数：9

共 44 条

[1]

Afsharimani N, 2019, B MATER SCI, V42, DOI 10.1007/s12034-018-1695-y

[2]

[Алексанин А.И. Alexanin A.I.], 2019, [Современные проблемы дистанционного зондирования Земли из космоса, Sovremennye problemy distantsionnogo zondirovaniya Zemli iz kosmosa], V16, P9, DOI 10.21046/2070-7401-2019-16-1-9-16

[3] Documenting Local Food Knowledge at Hindukush: Challenges and Opportunities [J].