Fast Classification Networks For Signal Processing

We present a generalization of the corner classification approach to training feedforward neural networks that allows rapid learning of nonbinary data. These generalized networks, called fast classification (FC) networks, are compared against backpropagation and radial basis function networks and are shown to have excellent performance for prediction of time series and pattern recognition. FC networks do not require iterative training and they can be used in many signal processing applications where fast, nonlinear filtering provides an advantage.     

Adaptive digital communication receivers

We present a tutorial review of adaptive signal processing applied to a digital communication receiver operating in a nonstationary environment. The topics covered include feedforward and recurrent neural networks and their applications to communication systems. The article also includes a discussion of turbo decoding, which involves a form of recurrent learning. It concludes with a discussion of the various aspects of adaptive signal processing as they relate to digital communication receivers.     

Connexionnisme,calcul, reconnaissance des formes et intelligence artificielle

This paper provides an introduction to the field of neural nets and associated learning procedures. Artificial intelligence and pattern recognition are taken in it as synonyms. Massively parallel, neural-like, networks are receiving increasing attention as a mechanism for expressing information processing models. The attempt to achieve human-like performances in the field of speech and image recognition justify this broadening of interest. Neural networks process information in new ways. The use of their properties appears to have promise for the development of solutions to problems that have intractable or unknown algorithms or are too computationally intense.     

Theory and applications of neural networks for industrial controlsystems

The theory and the applications of artificial neural networks, especially in a control field, are described. Recurrent networks and feedforward networks are discussed. Application to pattern recognition, information processing, design, planning, diagnosis, and control are examined. Hybrid systems using the neural networks, fuzzy sets, and artificial intelligence (AI) technologies are surveyed     

神经网络在通信中的应用

神经元网络的应用研究是近年来的一个研究热点。研究表明,在诸如自适应信号处理、最佳接收、纠错编码、压缩编码、模式识别、通信网等通信领域中,神经元网络可望得到广泛的应用。本文综述了几种神经元网络在通信中的应用研究工作。     

Learning pattern classification-a survey

Classical and recent results in statistical pattern recognition and learning theory are reviewed in a two-class pattern classification setting. This basic model best illustrates intuition and analysis techniques while still containing the essential features and serving as a prototype for many applications. Topics discussed include nearest neighbor, kernel, and histogram methods, Vapnik-Chervonenkis theory, and neural networks. The presentation and the large (though nonexhaustive) list of references is geared to provide a useful overview of this field for both specialists and nonspecialists     

Signal processing and pattern recognition with soft computing

We describe the overall role of soft computing (SC) in signal processing and pattern recognition (SPPR) with specific applications to biomedical engineering, geoscience for mining and civil engineering human interfaces, and image processing. Detection of characteristic points in an electrocardiogram to implement an advanced ECG analyzer is presented which is carried out using both conventional SPPR techniques and self-organizing neural networks. Successful technologies for monitoring a geostructure by supervised and self-organizing neural networks are described. Identification of a freehand drawing by a combination of fuzzy logic and neural networks is also described. Moreover, application of fuzzy logic to image segmentation is presented. Finally, innovation of SPPR using SC technologies is discussed     

深度卷积神经网络图像识别模型对抗鲁棒性技术综述

近年来,以卷积神经网络为代表的深度识别模型取得重要突破,不断刷新光学和SAR图像场景分类、目标检测、语义分割与变化检测等多项任务性能水平.然而深度识别模型以统计学习为主要特征,依赖大规模高质量训练数据,只能提供有限的可靠性能保证.深度卷积神经网络图像识别模型很容易被视觉不可感知的微小对抗扰动欺骗,给其在医疗、安防、自动...     

两个神经网络优化准则及其在无限制字符识别中的应用

本文根据信息理论提出了模式特征有效提取准则和稀疏有象准则两个新的神经网络优化准则，探讨了它们之间的相互关系，并给出了具体的能量函数形式和优化算法，设计了两个具有神经网络模型，根据上述地无限制手写体字符识别的初步应用中得到了很好的结果。     

Information theory, complexity and neural networks

Some of the main results in the mathematical evaluation of neural networks as information processing systems are discussed. The basic operation of feedback and feed-forward neural networks is described. Their memory capacity and computing power are considered. The concept of learning by example as it applies to neural networks is examined     

Neural networks for intelligent multimedia processing

This paper reviews key attributes of neural processing essential to intelligent multimedia processing (IMP). The objective is to show why neural networks (NNs) are a core technology for the following multimedia functionalities: (1) efficient representations for audio/visual information, (2) detection and classification techniques, (3) fusion of multimodal signals, and (4) multimodal conversion and synchronization. It also demonstrates how the adaptive NN technology presents a unified solution to a broad spectrum of multimedia applications. As substantiating evidence, representative examples where NNs are successfully applied to IMP applications are highlighted. The examples cover a broad range, including image visualization, tracking of moving objects, image/video segmentation, texture classification, face-object detection/recognition, audio classification, multimodal recognition, and multimodal lip reading     

基于GFMM神经网络的雷达目标识别方法

新型的无师训练（General Fuzzy Min—Max，GFMM）神经网络是一种具备无师训练聚类识别能力的新型神经网络，它继承了原有GFMM网络的特点，在网络的拓扑结构和算法方面进行了较大的改进，增加了能够进行自适应在线学习的能力。基于无师训练GFMM神经网络的雷达目标识别方法完整地实现了雷达目标特征学习和识别的一体化过程。在某型对海警戒雷达舰船目标识别仿真应用实验中的结果表明：文中的方法优于其他传统的神经网络目标识别方法，在雷达目标识别方面具有良好的适用性。     

基于多任务卷积神经网络的人脸识别技术研究

深度神经网络是目前计算机机器学习领域的一个关键技术,可应用于图像处理。其中,多任务卷积神经网络(Multi-task Convolutional Neural Network,MTCNN)是一种基于卷积神经网络的多任务人脸检测框架,这里采用MTCNN人脸检测模型代替传统的卷积神经网络,在深度学习框架TensorFlow上进行人脸识别。首先,在数据预处理阶段利用灰度化方法将图像集转变为灰度图,降低图像通道。其次,基于MTCNN构建人脸检测模型,并利用Softmax函数进行分类识别。最后,实验过程中选择不同迭代次数进行准确性对比,在模型趋于稳定的情况下,得到较高的准确性。     

基于MEMS谐振器硬件储备池计算的类脑信号处理方法

近年来兴起的人工神经网络由于具有较强的自学习适应性和并行信息处理能力,从而在信号处理领域显示出巨大潜力。储备池计算是一种由递归神经网络衍生而来的类脑神经形态计算范式,对随时间变化的连续信号具有非常好的分类和时序预测能力。本论文提出利用MEMS(Micro-Electro-Mechanical System)梁谐振器的非线性响应特征,设计并搭建了两种储备池计算的拓扑架构。此外,面向雷达信号处理中信号预测、图像识别、雷达信号特征分类和提取等应用需求,针对性地选择了NARMA(Nonlinear Auto Regressive Moving Average Equation of Order)预测任务、MNIST(Mixed National Institute of Standards and Technology)-手写数字图像识别、LFM(Linear frequency modulated)脉冲波形识别与特征提取等测试任务对论文所提两种不同储备池计算架构进行试验验证。同时,实验结果也充分展示了基于非线性MEMS谐振器的储备池计算硬件系统在雷达信号预测、分类与特征提取等应用领域中的应用...     

神经网络用于模式识别的几种主要方法及比较

神经网络的一个主要应用就是模式识别，本文主要讨论了用于模式识别的几种有代表性的神经网络类型及其学习或设计规则，对其优缺点进行了简要分析。     

BP网络的Matlab实现及应用研究

人工神经网络以其具有信息的分布存储、并行处理以及自学习能力等优点,已经在信息处理、模式识别、智能控制及系统建模等领域得到越来越广泛的应用。他的基于误差反向传播算法的多层前馈网络,即BP网络在非线性建模、函数逼近和模式识别中有广泛的应用,介绍了BP网络的基本原理,分析了Matlab人工神经网络工具箱中有关BP网络的工具函数,并给出了部分重要工具函数的实际应用。     

基于小波神经网络的与文本无关说话人识别方法研究

基于神经网络的说话人识别方法可以在一定程度上模仿人脑的功能,是说话人识别中的一种主要技术,但它通常难以确定隐层单元的数目,收敛速度慢,易于收敛到极小点。该文研究了一种用于说话人识别的小波神经网络模型,给出了网络结构和学习算法。采用Mel频率倒谱系数作为与文本无关的说话人识别的特征参数,并利用该模型进行了5个人的说话人识别实验,得到99.5%的识别率。实验结果表明,小波网络和传统的BP网络相比,训练速度和识别率都有了较大提高,具有良好的应用前景和进一步研究的价值。     

电子鼻信号处理方法综述

具有模仿人的嗅觉系统的电子鼻在过去十年中发展迅速,大部分成果开始商业应用,主要应用于食品和化妆品行业。用于电子鼻系统的信号处理方法主要包括静特征分析法和动态特征处理方法。静态特征分析法包括主成分分析法,判别函数分析法,类聚分析法和基于网络的多层感知器。动态特征分析法包括传统的参数法和非参数法,非参数法是借助于传统的系统识别方式及线性滤波器、时间序列神经网络系统。     

Gradient-based learning applied to document recognition

Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day