Title :
Shapes Extraction Method by Genetic Algorithm with Local Search Method
Author :
Matayoshi, Mitsukuni
Author_Institution :
Dept. of Ind. & Inf. Sci., Okinawa Int. Univ., Ginowan, Japan
Abstract :
Shapes extraction methods for getting two or more characters in convex hull or approximate convex hull are proposed in this paper. Proposed approaches use Genetic Algorithm (GA) with a improved local search or new local search method, which can get some characters as a shrink-wrapping from convex hull or approximate convex hull. The test problems are newly provided, which are made of the original problems in previous study. After obtaining convex hull or approximate convex hull, a local search method is applied to get the shape of objects in the hull. The experimental results show that proposed local search methods get the shape of objects with success.
Keywords :
feature extraction; genetic algorithms; object recognition; search problems; shape recognition; GA; approximate convex hull; genetic algorithm; local search method; object shape extraction method; Biological cells; Data mining; Genetic algorithms; Genetics; Search methods; Shape; Wrapping; computational geometry; convex hull; genetic algorithm; local search method; shape extraction; shrink-wrapping;
Conference_Titel :
Artificial Intelligence, Modelling and Simulation (AIMS), 2014 2nd International Conference on
Print_ISBN :
978-1-4799-7599-0
DOI :
10.1109/AIMS.2014.25