书写文字识别算法研究

介绍模式识别的基本原理和对书写文字字符的识别以及文字识别,阐述模式识别的过程、图像文件的格式,介绍基于住图技术的模式识别理论;讨论分类器设计过程中的特征空间优化设计问题、分类嚣设计准则、基本方法和判别函数的知识,给出模板匹配法识别图像的原理.     

隐空间中参数化直接鉴别分析及其应用

提出了一种新的非线性特征抽取方法——隐空间中参数化直接鉴别分析。其主要思想是利用一核函数将原始输入空间非线性变换到隐空间,针对在该隐空间中类内散布矩阵总是奇异等问题,利用参数化直接鉴别分析进行特征抽取。与现有的核特征抽取方法不同的是,该方法不需要核函数满足Mercer 定理,从而增加了核函数的选择范围。更为重要的是,由于在隐空间中采用了参数化直接鉴别分析,不仅保留了参数化直接鉴别分析的优点,而且有效地抽取了样本的非线性特征;在该方法中提出了一个更为合理的加权系数矩阵,提高了分类性能。在FERET人脸数据库子库上的实验结果验证了该方法的有效性。     

基于GRA-DDA加权耦合模型的煤层突出危险性评价

为了提高煤层突出危险性评价的效率和准确度,引入距离判别分析法(DDA),选用煤层瓦斯放散初速度、煤的坚固性系数、煤层瓦斯压力、煤体破坏类型和开采深度作为判别因子,并运用灰色关联分析(GRA)计算权重矩阵,建立了煤层突出危险性评价的加权距离耦合判别模型。将30个煤与瓦斯突出实例作为学习样本进行训练,建立相应的判别准则,经过训练后的模型误判率为0。利用该模型对10个煤与瓦斯突出实例进行评判,并将评判结果与单项指标法、BP神经网络法和未加权距离判别法进行对比分析,结果表明GRA-DDA耦合模型的评判结果准确,符合现场实际情况。     

无限最大间隔线性判别投影模型

针对具有多模分布结构的高维数据的分类问题,该文提出一种无限最大间隔线性判别投影(iMMLDP)模型。与现有全局投影方法不同,模型通过联合Dirichlet过程及最大间隔线性判别投影(MMLDP)模型将数据划分为若干个局部区域,并在每一个局部学习一个最大边界线性判别投影分类器。组合各局部分类器,实现全局非线性的投影与分类。iMMLDP模型利用贝叶斯框架联合建模,将聚类、投影及分类器进行联合学习,可以有效发掘数据的隐含结构信息,因而,可以较好地对非线性可分数据,尤其是具有多模分布特性数据进行分类。得益于非参数贝叶斯先验技术,可以有效避免模型选择问题,即局部区域划分数量。基于仿真数据集、公共数据集及雷达实测数据集验证了所提方法的有效性。     

基于PLS-LDA和拉曼光谱快速定性识别食用植物油

以6种食用油共计23个样本为分析对象,采用偏最小二乘线性判别分析法（PLS-LDA）和拉曼光谱进行单一种类（橄榄油、花生油和玉米油）食用油快速定性检测,通过自适应迭代惩罚最小二乘法（airPLS）对拉曼信号进行背景扣除,以及蒙特卡洛无信息变量消除法筛选波长变量,不但有效减少了波长点数,降低了建模运算量,而且提高了单一种类食用油的识别率,使得总体识别率均高于90%,并在此基础上进一步提出了采用PLS-LDA进行多种类食用油识别的检测流程。实验结果表明PLS-LDA在食用油定性识别检测中具有较好的应用前景和可行性,该方法也可为定性检测食品及农产品品质提供借鉴。      

无线传感网中多传感器特征融合算法研究

面向无线传感器网络在地面目标识别方面的应用需求,该文提出了一种基于改进局域判别基(Local Discriminant Bases, LDB)和二进制粒子群优化(Binary Particle Swarm Optimization, BPSO)方法的多传感器特征融合算法。利用新的基于概率密度估计的相对微分熵可分性测度来改进LDB,以提取目标信号的特征频段,然后分别利用一种改进的和一种全新的BPSO来实现特征融合。基于实地采集到的地面目标的声音和震动信号,仿真实验表明,该方法减少了所需分类器的数目,降低了特征维数,并在一定程度上提高了目标的正确识别率,具有实际的应用价值。     

基于GSVD的核不相关辨别子空间与雷达目标识别

该文提出了一种基于广义奇异值分解的核不相关辨别子空间算法,并将其用于高分辨距离像雷达目标识别。新算法结合广义奇异值分解与核方法的优点,有效地解决了传统方法面临的矩阵奇异问题,同时进一步改善了目标的类可分性。其次,依据Fisher准则导出了距离像总散度矩阵零空间中不含有有用辨别信息的结论。利用这一结论,可以在求解核不相关最优辨别矢量之前对各散度矩阵进行预降维,以减小后续运算的计算复杂度。对3类飞机目标实测数据的识别结果表明了所提方法的有效性。     

Medical Diagnosis System Based on Fast-weights Scheme

Clinical examination data often have the features of carrying vague information, missing data and incomplete examination records, which lead to higher probabilities of misdiagnosis. A variational recursive-discriminant joint model with fast weights (FWs) scheme is proposed. MIMIC-III data sets are trained and tested, and the results are used to diagnosing. Variational recurrent neural network (VRNN) with FWs can better obtain the temporal features with partly missing data, and discriminant neural network (DNN) is for decision. Moreover, layer regularization (LN) avoids the overflow of loss function and stabilize the dynamic parameters of each layer. For the simulations, 10 laboratory tests were selected to predict 10 diseases, 1600 samples and 400 samples were used for training and testing, respectively. The test accuracy of disease diagnosis without FWs is 72.55%, and that with FWs is 85.80%. Simulations reveal that the FWs mechanism can effectively optimize the system model, abstracting the features for diagnose, and significantly improve the accuracy of decision-making.     

Training pattern replication and weighted class allocation in artificial neural network classification

In some image classifications the importance of classes varies, and it is desirable to weight allocation to selected classes. Often the desire is to weight allocation in favour of classes that are abundant in the area represented by an image at the expense of the less abundant classes. If there is prior knowledge on the distribution of class occurrence, this weighting can be achieved with widely used statistical classifiers by setting appropriate a prioriprobabilities of class membership. With an artificial neural network, the incorporation of prior knowledge is more problematic. An approach to weight class allocation in an artificial neural network classification by replicating selected training patterns is presented. This investigation focuses on a series of classifications in which some classes were more abundant than others, but the same number of training cases were available for each class. By replicating the training patterns of abundant classes the representation of the abundant classes in the training set is increased, reflecting more closely the relative abundance of the classes in an image. Significant increases in classification accuracy were obtained by replicating the training patterns of abundant classes. Furthermore, in comparison against a discriminant analysis for the classification of synthetic aperture radar imagery, the results showed that training pattern replication could be used to weight class allocation with an effect similar to that of incorporating a prioriprobabilities of class membership into the discriminant analysis, and resulted in a significant 20.88%, increase in classification accuracy. This increase in classification accuracy was obtained without any new information, but was the result of making fuller use of what was available.     

Identification of adulterants in olive oils

The application of discriminant analysis for identifying and quantifying adulterants in extra virgin olive oils is presented. Three adulterants were used (sunflower oil, rapessed oil, and soybean oil) and were present in the range 5–95%. Near-infrared spectroscopy and principal components analysis were used to develop a discriminant analysis equation that could identify correctly the type of seed oil present in extra virgin olive oil in 90% of cases. Partial least squares analysis was used to develop a calibration equation that could predict the level of adulteration. Cross validation suggested that it was possible to measure the level of adulteration to an accuracy of ±0.9%. External validation of the derived calibation equation gave a standard error of performance of ±2.77%.