A Modified Pulse Coupled Neural Network with Anisotropic Synaptic Weight Matrix for Image Edge Detection
スポンサーリンク
概要
- 論文の詳細を見る
Pulse coupled neural network (PCNN) is a new type of artificial neural network specific for image processing applications. It is a single layer, two dimensional network with neurons which have 1 : 1 correspondence to the pixels of an input image. It is convenient to process the intensities and spatial locations of image pixels simultaneously by applying a PCNN. Therefore, we propose a modified PCNN with anisotropic synaptic weight matrix for image edge detection from the aspect of intensity similarities of pixels to their neighborhoods. By applying the anisotropic synaptic weight matrix, the interconnections are only established between the central neuron and the neighboring neurons corresponding to pixels with similar intensity values in a 3 by 3 neighborhood. Neurons corresponding to edge pixels and non-edge pixels will receive different input signal from the neighboring neurons. By setting appropriate threshold conditions, image step edges can be detected effectively. Comparing with conventional PCNN based edge detection methods, the proposed modified PCNN is much easier to control, and the optimal result can be achieved instantly after all neurons pulsed. Furthermore, the proposed method is shown to be able to distinguish the isolated pixels from step edge pixels better than derivative edge detectors.
- The Institute of Electronics, Information and Communication Engineersの論文
著者
-
Hu Jinglu
Graduate School Of Information Production And Systems Waseda Univ.
-
SHI Zhan
Graduate School of Information, Production and Systems, Waseda University
関連論文
- Elevator Group Supervisory Control System Using Genetic Network Programming with Macro Nodes and Reinforcement Learning
- A Double-Deck Elevator Group Supervisory Control System with Destination Floor Guidance System Using Genetic Network Programming
- Application of Universal Learning Networks to PV-Supplied DC Moter Drives
- 1C3-4 Benchmark Test of RasID-GA for Inequality/Equality Constrained Optimization
- Propagation and control of stochastic signals through universal learning networks
- A functions localized neural network with branch gates
- Hybrid Universal Learning Networks
- 2次元大津法とヒストグラム解析に基づく腎組織画像分割
- Genetic Network Programming with Reinforcement Learning and Its Application to Making Mobile Robot Behavior
- Association Rule Mining for Continuous Attributes using Genetic Network Programming
- KDI-Based Robust Fault Detection in Presence of Nonlinear Undermodeling
- A Hybrid Quasi-ARMAX Modeling Scheme for Identification of Nonlinear Systems
- Increasing Robustness of Binary-coded Genetic Algorithm
- An Incremental Learning of Neural Network with Multiplication Units for Function Approximation
- Enhancing the Generalization Ability of Neural Networks by Using Gram-Schmidt Orthogonalization Algorithm
- A Hybrid EDA for Protein Folding Based on HP Model
- An Adaptive Niching EDA with Balance Searching Based on Clustering Analysis
- Quasi-ARX wavelet network for SVR based nonlinear system identification
- Feature Subset Selection : A Correlation-Based SVM Filter Approach
- Contour Extraction of Glomeruli by Using Genetic Algorithm for Edge Patching
- A Two-Step Scheme for Polynomial NARX Model Identification Based on MOEA with Prescreening Process
- A New Learning Method Using Prior Information of Neural Networks
- Network Structure Oriented Evolutionary Model : Genetic Network Programming : Its Comparison with Genetic Programming
- A Neurofuzzy-Based Adaptive Predictor for Control of Nonlinear Systems
- Identification of Quasi-ARX Neurofuzzy Model with an SVR and GA Approach
- A Modified Pulse Coupled Neural Network with Anisotropic Synaptic Weight Matrix for Image Edge Detection
- A Quasi-ARX Neural Network with Switching Mechanism to Adaptive Control of Nonlinear Systems