A method for extracting outline using the genetic algorithm based on factors for perceptive grouping

• Published in: Proceedings 5th IEEE International Workshop on Robot and Human Communication. RO-MAN'96 TSUKUBA pp. 364 - 369
• Main Author: Saitoh, F.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 1996
• Subjects: Computer science; Data mining; Genetic algorithms; Genetic mutations; Humans; Image recognition; Joining processes; Noise shaping; Pixel; Shape
• DOI: 10.1109/ROMAN.1996.568864
• ISBN: 9780780332539; 0780332539

A method for extracting a global outline in a noisy image is proposed. The genetic algorithm is applied to extract a proper outline in an image which includes disconnected outline pixels and noisy pixels. Each individual in a population for the genetic algorithm has a sequence of outline pixel numbers as it genotype. The fitness of each individual is defined using factors represented by a closeness of pixels and a continuity of an outline. Individuals which have higher fitness survive to the next generation. Parents are selected in surviving individuals and their genes information is crosscovered to generate a child. Mutation is applied to improve variation of evolutionary processes. As the experimental results, the global proper outline is extracted using the evolutionary operations and rules in the genetic algorithm.