脈沖耦合神經(jīng)網(wǎng)絡(luò)及應(yīng)用

前言

There is no more complicated, advantaged and powerful device than themammalian primate cortical visual system for image processing in nature.The pulse-coupled neural network （PCNN） is inspired from the investigationof pulse synchronization within the mammalian visual cortex, and has beenwidely applied to image processing and pattern recognition.Visual cortex is the passage for brain to acquire information from eyesand a part of brain central nervous system. Several biological models basedon visual cortex were proposed through investigation of cat cortex and hadbeen applied to image processing.The PCNN emulates the mammalian visual cortex, which is supposed tobe one of the most efficient image processing methods. The output of thePCNN is a series of pulse images which represent the fundamental featuresof original stimulus, such as edge, texture, and segment. Neurons receiveinputs from other neurons through synapses and are fired synchronously incertain regions, that is why the PCNN can be applied to image segmentation,smoothing, and coding. Another important feature of the PCNN is that thepulse images are able to be characterized to a unique invariant signature forthe image retrieval.This book analyzes the PCNN in detail and presents some special appli-cations and corresponding results based on our own researches.Contributions of the book have come from Hongjuan Zhang, RongchangZhao, Maojun Su, Dongmei Lin, Xiaojun Li, Guanzhu Xu, Xin Wang, Za-ifeng Zhang, Xiaowen Feng, Haibo Deng, Li Liu, Xiaozhe Xu, ChunliangQi, Chenghu Wu, Fei Shi, Zhibai Qian, Qing Liu, Min Yuan, Jiuwen Zhang,Yingjie Liu, Xiaolei Chen, and our graduate students at Circuit and SystemResearch Institute of Lanzhou University.

內(nèi)容概要

Applications of Pulse-Coupled Neural Networks explores the fields of image processing, including image filtering, image segmentation, image fusion, image coding, image retrieval, and biometric recognition, and the role of pulse-coupled neural networks in these fields. 　　This book is intended for researchers and graduate students in artificial intelligence, pattern recognition, electronic engineering, and computer science.

書籍目錄

Chapter 1 Pulse-Coupled Neural Networks　1.1　Linking Field Model　1.2　PCNN　1.3　Modified PCNN　　1.3.1　Intersection Cortical Model 　　1.3.2　Spiking Cortical Model　　1.3.3　Multi-channel PCNN　Summary 　References Chapter 2　Image Filtering　2.1　Traditional Filters　　2.1.1　Mean Filte　　2.1.2　Median Filte　　2.1.3　Morphological Filter　　2.1.4　Wiener Filter　2.2　Impulse Noise Filtering 　　2.2.1　Description of Algorithm Ⅰ　　2.2.2　Description of Algorithm Ⅱ　　2.2.3　Experimental Results and Analysis　2.3　Gaussian Noise Filtering 　　2.3.1　PCNNNI and Time Matrix　　2.3.2　Description of Algorithm Ⅲ　　2.3.3　Experimental Results and Analysis　Summary　ReferencesChapter 3　Image Segmentation　3.1　Traditional Methods and Evaluation Criteria　　3.1.1　Image Segmentation Using Arithmetic Mean　　3.1.2　Image Segmentation Using Entropy and Histogram　　3.1.3　Image Segmentation Using Maximum Between-cluster Variance　　3.1.4　Objective Evaluation Criteria　3.2　Image Segmentation Using PCNN and Entropy　3.3　Image Segmentation Using Simplified PCNN and GA　　3.3.1　Simplified PCNN Model　　3.3.2　Design of Application Scheme of GA　　3.3.3　Flow of Algorithm　　3.3.4　Experimental Results and Analysis　Summary　ReferencesChapter 4　Image Coding　4.1　Irregular Segmented Region Coding　　4.1.1　Coding of Contours Using Chain Code　　4.1.2　Basic Theories on Orthogonality　　4.1.3　Orthonormalizing Process of Basis Functions　　4.1.4　ISRC Coding and Decoding Framework　4.2　Irregular Segmented Region Coding Based on PCNN　　4.2.1　Segmentation Method　　4.2.2　Experimental Results and Analysis　Summary　ReferencesChapter 5　Image Enhancement　5.1　Image Enhancement　　5.1.1　Image Enhancement in Spatial Domain　　5.1.2　Image Enhancement in Frequency Domain　　5.1.3　Histogram Equalization　5.2　PCNN Time Matrix　　5.2.1　Human Visual Characteristics　　5.2.2　PCNN and Human Visual Characteristics　　5.2.3　PCNN Time Matrix　5.3　Modified PCNN Model　5.4　Image Enhancement Using PCNN Time Matrix　5.5　Color Image Enhancement Using PCNN　Summary　References Chapter 6　Image FusionChapter 7　Feature ExtractionChapter 8　Combinatorial OptimizationChapter 9　FPGA Implementation of PCNN AlgorithmIndex

章節(jié)摘錄

插圖：（1） If all the values in the structuring element are positive, the outputimage tends to be brighter than the input.（2） Dark elements within the image are reduced or eliminated, dependingon how their shapes relate to the structuring element used.The degree of these effects depends greatly on the shape and values withinthe structuring element and the details within the image itself.Grayscale erosion is defined as the minimum of the difference of a lo-cal region of an image and a grayscale mask. The shape of the input mask（known as the structuring element, SE） is generally chosen to emphasize orde-emphasize elements in the image. It is used to smooth small light regions.The general effects of performing erosion on a grayscale image are asfollows:（1） If all the values in the structuring element are positive, the outputimage tends to be darker than the input.（2） Light elements within the image are reduced or eliminated, dependingon how their shapes relate to the structuring element used.The degree of these effects depends greatly on the shape and values withinthe structuring element and the details within the image itself.Grayscale morphological opening of an image is defined as the dilation ofthe erosion of the image. The result is the reduction of small positive regionswithin the image. Grayscale morphological closing of an image is defined asthe erosion of the dilation of the image. The result is the reduction of smallnegative regions within the image.

編輯推薦

《脈沖耦合神經(jīng)網(wǎng)絡(luò)及應(yīng)用(國內(nèi)英文版)》是由高等教育出版社出版。

圖書封面

評論、評分、閱讀與下載

---

    脈沖耦合神經(jīng)網(wǎng)絡(luò)及應(yīng)用 PDF格式下載

---