A robust method for estimating projective transformations using genetic algorithms

Song, Ran and Szymanski, John (2007) A robust method for estimating projective transformations using genetic algorithms. In: IADIS Computer Graphics and Visualization, 5 - 7 July 2007, Lisbon, Portugal.

Full text not available from this repository.

Item Type:Conference or Workshop contribution (Paper)
Item Status:Live Archive


This paper presents a robust method which provides quantitative estimates of the projective transformations between two successive overlapping images using genetic algorithms. In this method, roulette selection and total arithmetic crossover are applied based on real number encoding. Then an adaptive mutation operator is used to preserve the best solutions. The experimental results show that the normalized registration error of the final solution exhibits a significant improvement over those obtained by direct search approaches to such problems. Also, in contrast to other popular approaches such as the least-squares and Levenberg-Marquardt algorithms, this algorithm can escape from local extrema and can potentially realize the global optimum.

Keywords:Genetic Algorithms
Subjects:G Mathematical and Computer Sciences > G740 Computer Vision
Divisions:College of Science > School of Computer Science
Related URLs:
ID Code:13301
Deposited On:08 Feb 2014 22:48

Repository Staff Only: item control page