Title:	REACTION-DIFFUSION ALGORITHM DESIGNED FOR IMAGE EDGE DETECTION AND ITS PERFORMANCE EVALUATION
Author(s):	Atsushi Nomura, Koichi Okada and Yoshiki Mizukami
ISBN:	978-989-8533-79-1
Editors:	Katherine Blashki and Yingcai Xiao
Year:	2018
Edition:	Single
Keywords:	Edge Detection, Reaction-Diffusion, Fitzhugh-Nagumo Type Model, Initial Condition, Edge Thinning
Type:	Full Paper
First Page:	269
Last Page:	276
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	The present paper proposes image edge detection algorithm utilizing a reaction-diffusion network. The network consists of two-dimensionally coupled FitzHugh-Nagumo type neurons, whose behavior is described by a set of time-evolving partial differential equations. The neurons’ non-linearity and their interconnection give the image edges through an iterative computation. The proposed algorithm differs from its previously-proposed one (Nomura et al., 2016) on the pre-processing for initializing the neurons’ state and the post-processing for thinning edges. The performance of the proposed algorithm is discussed with three image data-sets and two error measures, in comparison with performance of two other algorithms including Canny’s edge detection method.
	Go Back