PARTICLE SWARM OPTIMIZATION FOR TEMPLATE MATCHING
|
Accession number;07A0187218
|
| Title;PARTICLE SWARM OPTIMIZATION FOR TEMPLATE MATCHING |
| Author;
ANDO HIROSHI
(Yokohama National Univ., Kanagawa, Jpn)
NAGAO TOMOHARU
(Yokohama National Univ., Kanagawa, Jpn)
|
Journal Title;IEIC Technical Report (Institute of Electronics, Information and Communication Engineers)
|
Journal Code:S0532B
|
ISSN:0913-5685
|
|
VOL.106;NO.448(IE2006 138-184);PAGE.237-242(2007)
|
| Figure&Table&Reference; |
| Pub. Country;Japan |
| Language;English |
| Abstract;In this paper we propose an approach, which uses Particle Swarm Optimization (PSO) for template matching problems. Template matching is an important technique to investigate similarities between two patterns, predefined patterns (template) and target patterns. Template matching is also useful for detection of an object by determining angle, position and magnification ratio in an image. For detection of a two dimensional object, four parameters (angle, x-coordinate, y-coordinate and magnification ratio) have to be optimized in template matching. However, the search space consisting of these four parameters is so huge that it takes a lot of calculations to investigate all searching points. PSO is one of the new optimization algorithms which is categorized as an evolutionary algorithm, and based on an analogy of the movement of flight of a flock of birds. Each bird (called a particle in PSO) represents the searching point in search space. In this research, each particle has four dimensional vectors and updates either two known PSO algorithms (inertia weight approach or constriction coefficient approach). We apply PSO to the template matching problem, and compare the PSO based algorithm to the simple Genetic Algorithm (GA) in two kinds of two dimensional images. Experimental results show that the PSO based algorithm is able to obtain accurate results and locate an optimal solution significantly faster than GA. Besides, the PSO approaches outperform in terms of processing cost. (author abst.) |
|
|
|
Related Articles;
|
|
