基于卷积神经网络的印刷套准识别方法

目的 针对目前印刷套准识别方法依赖于经验人工设计特征提取的问题,提出一种不需要人工提取图像特征的卷积神经网络模型,实现印刷套准状态的识别.方法 采用图像增强技术实现不均衡训练集的均衡化,增加训练集图像的数量,提高模型的识别准确率.设计基于AlexNet网络结构的印刷套准识别模型的结构参数,分析批处理样本数量和基础学习率对模型性能的影响规律.结果 文中方法获得的总印刷套准识别准确率为0.9860,召回率为1.0000,分类准确率几何平均数为0.9869.结论 文中方法能自动提取图像特征,不依赖于人工设计的特征提取方法.在构造的数据集上,文中方法的分类性能优于实验中的支持向量机方法.     

基于颜色矢量角度直方图与DCT压缩的鲁棒哈希算法

目的为了改善哈希序列对篡改内容的正确识别率和鲁棒性,提出基于颜色矢量角度直方图耦合离散余弦变换压缩的鲁棒哈希算法。方法结合插值运算和Gaussian滤波器,对图像实施预处理,使其对于任意的认证目标均可产生一个固定长度的哈希序列;引入极坐标变换LPT(Log-polar transform),对规范尺寸图像完成坐标变换,输出二次图像;借助SVD(SingularValueDecompostion)方法来处理二次图像,输出全局鲁棒特征;提取规范尺寸图像中所有像素分量的颜色矢量角度,并计算其对应的直方图,形成色彩感知特征;将全局与色彩感知特征组合,将其视为1D哈希序列;基于离散余弦变换DCT(Discrete Cosine Transform),对1D哈希序列实施压缩,输出一组交流系数;引入Logistic映射,根据其输出的混沌序列来设计差异加密方法,对交流系数实施扩散,从而形成目标哈希序列。基于l2范数距离和优化的识别阈值,对待检测图像内容的真伪实施判别。结果试验结果显示,相对于已有的哈希机制而言,所提方案具备更强的鲁棒性,对各类几何攻击均有更高的检测准确度。结论所提哈希方法具有较高的鲁棒性与识别正确率,对商标检索、信息防伪等领域具有较好的参考价值。     

Multi classifier-based score level fusion of multi-modal biometric recognition and its application to remote biometrics authentication

Biometric recognition has become a common and reliable way to authenticate the identity of a person. Multimodal biometrics has become an interest of areas for researches in the recent past as it provides more reliability and accuracy. In multimodal biometric recognition, score level fusion has been a very promising approach to improve the overall system's accuracy. In this paper, score level fusion is carried out using three categories of classifiers like, rule classifier (fuzzy classifier), lazy classifier (Naïve Bayes) and learning classifiers (ABC-NN). These three classifiers have their own advantages and disadvantages so the hybridization of classifiers leads to provide overall improvements. The proposed technique consists of three modules, namely processing module, classifier module and combination module. Finally, the proposed fusion method is applied to remote biometric authentication. The implementation is carried out using MATLAB and the evaluation metrics employed are False Acceptance Rate (FAR), False Rejection Rate (FRR) and accuracy. The proposed technique is also compared with other techniques and by employing various combinations of modalities. From the results, we can observe that the proposed technique has achieved better accuracy value and Receiver Operating Characteristic (ROC) curves when compared to other techniques. The proposed technique reached maximum accuracy of having 95% and shows the effectiveness of the proposed technique.     

基于支持向量机的相位敏感光时域反射仪研究

为了解决实际环境中振动事件易误报的问题,在基于相位敏感光时域反射仪的分布式光纤振动传感系统中,引入了一种融合小波能量谱和支持向量机(SVM)的模式识别方法。首先,利用小波能量谱分析方法,设定最优分解层为5层,并从原始信号中提取出特征向量;然后利用支持向量机的“一对一法”多分类策略对振动事件进行识别分类。考虑到实际环境因素的影响,对沿光纤行走、敲击光纤以及沿光纤慢跑3种模式的振动进行了检测试验;最后,采用准确率、精确率、召回率和F值来综合评价支持向量机分类器的性能。实验结果表明:该模式识别方法实现了84.9%的振动事件分类准确率。     

On systems with shared resources and optimal switching strategies

Simple series systems of identical components with spare parts are considered. It is shown that the cumulative distribution function of a system failure time tends to be a step function as the number of components increases and resources can be shared. An example of ‘continuous resources’ is also described. The time-sharing strategy for standby systems is investigated. It is proved that an optimal rule for a system of standby components with increasing failure rates is the single switching performed at a=t/2, where t is the mission time.     

Illumination normalization using independent component analysis and filtering

In this work, we separate the illumination and reflectance components of a single input image which is non-uniformly illuminated. Considering the input image and its blurred version as two different combinations of illumination and reflectance components, we use the conventional independent component analysis (ICA) to separate these two components. The separated reflectance component, which is an illumination normalized version of the input image, can then be used as an effective pre-processed (illumination normalized) image for different computer vision tasks e.g. face recognition. To this end, we present simulation results to show that our proposed pre-processing method called illumination normalization using ICA increases the accuracy rate of several baseline face recognition systems (FRSs). The proposed method showed improved performance of baseline FRSs when using the Extended Yale-B databases.     

Non-zero order joint transform correlator with an optimized real-valued reference function

Abstract

In order to achieve distortion invariant image recognition, the design of the non-zero order joint transform correlator with minimum average cross-correlation energy has been studied. In this work, we prove that the optimal reference function is real-valued if the desired cross-correlation peak is also set to be real for every image in a training set. The recognition effectiveness is demonstrated by an example, where the genuine seal-imprint can be detected as compared with the invalid, although very similar one.     

Color Edge Detection Using Multidirectional Sobel Filter and Fuzzy Fusion

A new model is proposed in this paper on color edge detection that uses the second derivative operators and data fusion mechanism. The second-order neighborhood shows the connection between the current pixel and the surroundings of this pixel. This connection is for each RGB component color of the input image. Once the image edges are detected for the three primary colors: red, green, and blue, these colors are merged using the combination rule. Then, the final decision is applied to obtain the segmentation. This process allows different data sources to be combined, which is essential to improve the image information quality and have an optimal image segmentation. Finally, the segmentation results of the proposed model are validated. Moreover, the classification accuracy of the tested data is assessed, and a comparison with other current models is conducted. The comparison results show that the proposed model outperforms the existing models in image segmentation.     

Automatic transcription of continuous speech into syllable-like units for Indian languages

The focus of this paper is to automatically segment and label continuous speech signal into syllable-like units for Indian languages. In this approach, the continuous speech signal is first automatically segmented into syllable-like units using group delay based algorithm. Similar syllable segments are then grouped together using an unsupervised and incremental training (UIT) technique. Isolated style HMM models are generated for each of the clusters during training. During testing, the speech signal is segmented into syllable-like units which are then tested against the HMMs obtained during training. This results in a syllable recognition performance of 42·6% and 39·94% for Tamil and Telugu. A new feature extraction technique that uses features extracted from multiple frame sizes and frame rates during both training and testing is explored for the syllable recognition task. This results in a recognition performance of 48·7% and 45·36%, for Tamil and Telugu respectively. The performance of segmentation followed by labelling is superior to that of a flat start syllable recogniser (27·8%and 28·8%for Tamil and Telugu respectively).     

Optimizing water supply and hydropower reservoir operation rule curves: An imperialist competitive algorithm approach

Efficient reservoir management requires the implementation of generalized optimal operating policies that manage storage volumes and releases while optimizing a single objective or multiple objectives. Reservoir operating rules stipulate the actions that should be taken under the current state of the system. This study develops a set of piecewise linear operating rule curves for water supply and hydropower reservoirs, employing an imperialist competitive algorithm in a parameterization–simulation–optimization approach. The adaptive penalty method is used for constraint handling and proved to work efficiently in the proposed scheme. Its performance is tested deriving an operation rule for the Dez reservoir in Iran. The proposed modelling scheme converged to near-optimal solutions efficiently in the case examples. It was shown that the proposed optimum piecewise linear rule may perform quite well in reservoir operation optimization as the operating period extends from very short to fairly long periods.     

C/C复合材料组分含量的图像分析方法

C/C复合材料的组分含量测量是进行性能分析和改进工艺的有效手段。本文作者在分析化学气相渗透工艺(CVI)制备的纯净组织C/C复合材料偏振光显微图像特点的基础上, 基于模式识别原理提出了一种自适应多目标图像分割方法。根据最大类间方差准则、 采用改进的Otsu算法, 该系统自动计算孔隙、 纤维和热解炭各相间的最佳分割阈值。实验结果表明, 该方法不受C/C复合材料组分含量和组分分布形式的影响, 分割质量满足定量测量的要求。      

图片卫士:一个自动成人图像识别系统

设计并实现了一个自动识别成人图像识别系统“图片卫士”。图片卫士采用3层识别框架,利用肤色、纹理、图像视觉特征分层逐级识别成人图像。为了可靠地检测到图像中的肤色区域,提出了一种新的自适应统计肤色模型。在肤色检测基础上,通过皮肤纹理验证过程,图像中的人体皮肤区域被准确地分割出来。基于图像中皮肤区域,提取9个经验特征来表示图像内容,并采用AdaBoost算法构造一个总体分类器进行图像分类,识别正常图像和成人图像。在算法评估中,建立了一个78205幅图像的测试集,其中59885幅为正常图像,18320幅为成人图像。图片卫士显示了良好的系统性能,具有成人图像88．5％的识别率,正常图像92．5％的识别率。在PentiumⅣ1．5GHz的个人计算机上,图片卫士的平均处理速度为正常图像每秒5．6幅和成人图像每秒1．9幅。图片卫士可以应用在个人计算机或网络传输中,实时监控和过滤成人图像,还可以为网络安全等应用提供技术支持。     

基于机器视觉的金属丝网缺陷识别方法

目的 提升金属丝网的检测效率与检测精度。方法 提出一种应用于金属丝网表面缺陷识别的EfficientNetV2改进网络,首先更改了网络的骨干结构,在特征提取模块前后分别引入通道拆分与通道转换等操作,以增大网络容量,提高特征利用率;其次重新设计网络的分类器,通过对提取的高级语义信息进行逐层分步压缩,以减小特征损失,提高分类精度;最后搭建图像采集系统,构造金属丝网缺陷数据集。结果 实验结果表明,文中改进的网络模型在数据集上的准确率、精确度和特异度分别达到99.43%、99.42%和99.88%,图像识别耗时为27.5ms,增强了缺陷识别效果。结论 该方法具有较高的准确率,在金属丝网缺陷检测上具有较好的实用性,也可为其他类似产品的缺陷检测提供参考。     

Recognition and Classification of Pomegranate Leaves Diseases by Image Processing and Machine Learning Techniques

Disease recognition in plants is one of the essential problems in agricultural image processing. This article focuses on designing a framework that can recognize and classify diseases on pomegranate plants exactly. The framework utilizes image processing techniques such as image acquisition, image resizing, image enhancement, image segmentation, ROI extraction (region of interest), and feature extraction. An image dataset related to pomegranate leaf disease is utilized to implement the framework, divided into a training set and a test set. In the implementation process, techniques such as image enhancement and image segmentation are primarily used for identifying ROI and features. An image classification will then be implemented by combining a supervised learning model with a support vector machine. The proposed framework is developed based on MATLAB with a graphical user interface. According to the experimental results, the proposed framework can achieve 98.39% accuracy for classifying diseased and healthy leaves. Moreover, the framework can achieve an accuracy of 98.07% for classifying diseases on pomegranate leaves.     

基于支持向量数据描述的印刷图像套准状态检测方法

目的 针对不均衡的印刷标志图像训练集构建的二分类模型,对少类的印刷套不准图像识别准确率低的问题,研究不均衡印刷标志图像套准状态的单分类模型识别方法。方法 提出支持向量数据描述方法,实现多类的印刷套准图像和少类的印刷套不准图像的准确识别。采用多类的印刷套准图像训练支持向量数据描述,构建模型。采用网格寻优方法和交叉验证方法确定模型的最佳参数 和 。利用模型对印刷标志图像套准状态进行识别。结果 采用文中提出的支持向量数据描述方法,对印刷标志图像套准状态识别获得的总体识别率a为0.9500,印刷套准图像和印刷套不准图像识别准确率的几何平均数Gmean为0.9513。结论 文中提出的方法获得的总体识别率a和识别率的几何平均数Gmean要优于实验中的其他方法。     

MATHEMATICAL ANALYSIS OF FACILITY-COMMODITY SCHEDULING PROBLEMS

A number of operational situations exist in which certain facilities are available and where a number of commodities must be processed on some or all of these facilities. The paper describes an algorithm to generate schedules which are near optimal or optimal with respect to the total processing time of all the commodities, the idle lime of facilities and production rate. Thus, these schedules are characterized by near minimal or minimal total processing time and idle time of facilities and near maximal or maximal production rate. Usually this algorithm does not result in the desired schedule after the first application; it is therefore proposed to generate a set “D” of schedules from which the desired schedule can be selected. A decision rule determines the optimal number of elements belonging to set D

In order to justify the concept of the algorithm for the determination of the schedules mentioned above, an analysis is given of the decision tree associated with the sequencing model in terms of the probabilities related to the nodes in the decision tree.     

An Intelligent Two Phase Fuzzy Decision Tree Based Clustering Model for Design of Computer Aided Detection/Diagnosis (CADe/CADx) System

Automated computer aided detection/diagnosis (CADe/CADx) system plays a key role in decision making process and work as a recommender system for researchers. Nowadays, CADe/CADx systems are getting popular due to their strong ability to extract complex relations present in unprocessed data set. In this research article, we present an intelligent two phase classification model for the design of CADe/CADx system. In order to design an intelligent CAD system, the primary challenge lies in identifying important attributes. The presence of irrelevant and redundant attributes in the data can have adverse impact over classifier accuracy. An efficient dimensionality reduction technique aims at achieving lower computational cost with reduced storage requirement by choosing problem specific relevant or significant attributes. The secondary challenge is to provide unambiguous and comprehensible rule base for accurate predictions. The contribution of work can be stated twofold: first, to attain reasonably good classification accuracy with possible speed up, linear discriminant analysis and some popular correlation coefficients (Fisher, Phi and Point bi-serial) are being used to identify significant attributes. Second, to generate comprehensible and understandable rule set a fuzzy decision tree based clustering approach is used. The performance of proposed model is verified on twelve famous UCI data sets.     

面向不均衡训练集的印刷图像套准状态检测方法

目的针对不均衡的印刷图像套准状态检测中存在的印刷套不准图像识别准确率低的问题,研究不均衡印刷图像训练集的预处理方法。方法提出不均衡印刷图像训练集数据的集成采样预处理方法。支持向量机先将不均衡的训练集数据分为支持向量和非支持向量,然后过采集少类样本(即印刷套不准图像)中的支持向量,欠采集多类样本(即印刷套准图像)中的非支持向量,实现训练集数据的均衡化。最后采用预处理后的均衡训练集对支持向量机模型进行训练,并优化模型参数。结果采用文中提出的集成采样方法对不均衡训练集预处理后获得支持向量机模型,通过对印刷图像套准状态进行识别,获得的少类样本识别率a+为0.9375,识别准确率几何平均数Gmean为0.9437,F测度为0.9574。结论文中提出方法获得的印刷套不准图像识别准确率a+,Gmean和F测度均优于实验中的其他方法。     

基于特征选择和GWO-KELM的鸟声识别算法

针对鸟声识别算法中提取特征单一、分类准确率低等问题,提出一种基于混合特征选择和灰狼算法优化核极限学习机的鸟声识别方法。首先从鸟声数据中提取大规模声学特征集ComParE,其次计算每个特征的F_score并进行排序,然后以广义顺序向前浮动搜索(Generalized Sequential Forward Floating Search, GSFFS)为搜索策略,特征子集在核极限学习机(Kernel Limit Learning Machine, KELM)上十折交叉验证的正确率,作为特征选择标准进行特征选择,得到适用于鸟声识别的特征子集,最后通过灰狼算法(Grey Wolf Optimizer, GWO)选择最优KELM参数识别鸟声。在柏林自然科学博物馆鸟声数据库中进行实验,该方法在60类鸟声识别平均正确率和F1-score达到94.45%和92.29%。结果表明,该方法相较于传统自行设计提取的单一特征集具有更高的识别精度,GWO-KELM模型比网格搜索方式更易找到全局最优值。     

Scalable Skin Lesion Multi-Classification Recognition System

Skin lesion recognition is an important challenge in the medical field. In this paper, we have implemented an intelligent classification system based on convolutional neural network. First of all, this system can classify whether the input image is a dermascopic image with an accuracy of 99%. And then diagnose the dermoscopic image and the non-skin mirror image separately. Due to the limitation of the data, we can only realize the recognition of vitiligo by non-skin mirror. We propose a vitiligo recognition based on the probability average of three structurally identical CNN models. The method is more efficient and robust than the traditional RGB color space-based image recognition method. For the dermoscopic classification model, we were able to classify 7 skin lesions, use weighted optimization to overcome the unbalanced data set, and greatly improve the sensitivity of the model by means of model fusion. The optimization and expansion of the system depend on the increase of database.