One-dimensional reaction-diffusion systems in diffusive environment: emergence of long-range inhibition and its application to visual information processing

2016 Fiscal Year Final Research Report

• Project/Area Number: 26330276
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Soft computing
• Research Institution: Yamaguchi University
• Principal Investigator: Nomura Atsushi 山口大学, 教育学部, 教授 (40264973)
• Co-Investigator(Renkei-kenkyūsha): ICHIKAWA Makoto 千葉大学, 文学部, 教授 (10294654) ; SAKURAI Tatsunari 千葉大学, 理学研究科, 准教授 (60353322)
• Project Period (FY): 2014-04-01 – 2017-03-31
• Keywords: 反応拡散系 / 結合興奮系 / FitzHugh-Nagumo / 長距離抑制 / 視覚情報処理 / ノイズ除去 / 画像符号化 / エッジ検出

Outline of Final Research Achievements

This research work proposed algorithms of visual information processing: image noise reduction, intensity coding and edge detection. The algorithms utilize a system of coupled FitzHugh-Nagumo type nonlinear reaction elements. Each element has an internal state governed by two ordinary differential equations. The elements are placed on a two- or three-dimensional grid, and are coupled to their nearest neighboring elements under long range inhibition. By utilizing the grid, we successfully proposed the noise-reduction algorithm with a three-dimensional grid, the image intensity coding algorithm by introducing additive Gaussian noise into initial conditions, and the edge detection algorithm with external stimuli of image gradients.

Free Research Field

視覚情報処理