关键词： 薄层   河道砂体   调谐厚度   波峰/波谷比   相对时差   模式识别  

摘要： 薄河道砂体纵向叠置不规则、横向连续性差,存在人工解释工作量大且准确度不高的问题.在分析垂直河道的折线时发现,研究区薄砂体储层呈现出上面波谷、下面波峰的"双峰"特征,且波峰和波谷的时差趋于一定值,波峰和波谷的幅值也非常接近.基于以上反射特征或者说薄河道砂体的反射模式,研发了基于模式识别的薄砂体特征提取及表征方法:(1)将顶底时差、波峰/波谷振幅比、波峰最大振幅、波谷最小振幅4个参数,作为识别薄砂体的基本参数;(2)数据收索策略,考虑了时窗大小、滑动步长及层位约束3个控制因素;(3)依据搜索结果,还用基于调谐原理的相对振幅法计算了砂体的厚度.理论模型测试和实际资料应用表明,该方法可高效、自动地识别出整个三维工区的薄砂体并计算出砂体厚度,为识别和表征复杂河流相储层提供了一种新的途径.

关键词： 高光谱技术   酸奶   模式识别   菌落计数   益生菌  

摘要： 采用高光谱技术结合平板菌落法快速表征益生菌酸奶中常见的发酵菌种(保加利亚乳杆菌、嗜热链球菌)和益生菌种(干酪乳杆菌、嗜酸乳杆菌、植物乳杆菌)的光谱差异,并结合不同模式识别方法对混合菌种发酵酸奶中每种益生菌的数量进行同步计数,最后与传统计数法得到的菌落数量结果进行对比,通过2种计数结果的差异性分析验证高光谱菌落计数的可行性。结果表明,当主成分数为9时,最小二乘支持向量机模型对不同种类菌落识别效果均优于K-最近邻法和误差反向传播神经网络模型,其中校正集识别率为99.20%,预测集识别率为93.33%,为最佳计数模型,与传统计数法并无显著差异(P>0.05),验证高光谱技术应用在混合菌种发酵酸奶中对每种益生菌同步计数的可行性,解决传统计数法无法同步计数混合菌种酸奶中各益生菌数量的缺陷问题,为快速无损检测酸奶品质与保健活性提供检测依据。

关键词： 雷达工作模式   雷达短语   粒子群寻优   离散过程神经网络  

摘要： 针对参数交叠严重环境下的雷达工作模式识别问题,提出一种基于粒子群优化算法(PSO)的离散过程神经网络(process neural network, DPNN)的识别方法。方法依据整个雷达信号脉冲序列的时序变化规律进行识别,首先对雷达信号句法建模并提取雷达短语作为工作模式的特征描述,然后运用PSO有监督训练合适的DPNN网络结构,最后运用训练完成的DPNN识别未知雷达短语的工作模式。对比仿真结果表明,信号参数测量误差10%时识别率为97%,较传统识别方法提高30%,方法在参数交叠严重的情况下的工作模式识别率和抗误差性能提升明显。

关键词： Behavior   Software   Manufacturing   Signal processing   Artificial intelligence   Scrap   Machine learning   Manufacturers   Pattern recognition   Digital signatures   Quality   New technology  

摘要： The article offers the information on the Use of Patter Recognition to Improve Industrial Operations. It mentions about the Machine learning through pattern recognition can be a faster, less expensive approach to monitoring manufacturing; and Manufacturers use a variety of methodologies to monitor operations. Foremost is statistical process control (SPC), except that SPC parameters need to be known before they can be monitored and acted upon.

关键词： Digital mammogram   Microcalcification   Pattern recognition   Binning   Reflection coefficient   Mass density   3D interpolation   Size calculation of microcalcification  

摘要： Breast cancer is one of the life-threatening cancers occurring in women. In recent years, from the surveys provided by various medical organizations, it has become clear that the mortality rate of females is increasing owing to the late detection of breast cancer. Therefore, an automated algorithm is needed to identify the early occurrence of microcalcification, which would assist radiologists and physicians in reducing the false predictions via image processing techniques. In this work, we propose a new algorithm to detect the pattern of a microcalcification by calculating its physical characteristics. The considered physical characteristics are the reflection coefficient and mass density of the binned digital mammogram image. The calculation of physical characteristics doubly confirms the presence of malignant microcalcification. Subsequently, by interpolating the physical characteristics via thresholding and mapping techniques, a three-dimensional (3D) projection of the region of interest (RoI) is obtained in terms of the distance in millimeter. The size of a microcalcification is determined using this 3D-projected view. This algorithm is verified with 100 abnormal mammogram images showing microcalcification and 10 normal mammogram images. In addition to the size calculation, the proposed algorithm acts as a good classifier that is used to classify the considered input image as normal or abnormal with the help of only two physical characteristics. This proposed algorithm exhibits a classification accuracy of 99%.

关键词： LC-MS   Total Ion Current Chromatogram (TICC)   Noise  

摘要： Total ion current chromatograms (TICCs) generated by liquid chromatography-mass spectrometry (LC-MS) are prone to noise from chemical and electronic sources. This noise can severely impact the detection of analytes in a mixture. Recently, we introduced a new variable selection tool based on Pattern Recognition Entropy (PRE) that selects good quality (high signal-to-noise ratio) mass chromatograms from an LC-MS dataset and thereby creates a reduced TICC with low noise and a flat background (J. Chrom. A. 2018, 1558, 21-28). PRE, which is based on Shannon's entropy, was shown to be a straightforward and powerful shape recognition tool for this problem. However, while the chromatographic signals in the reduced TICC from PRE were well resolved, some noise remained in the TICC, which suggested that the algorithm had selected some false positives, i.e., poor quality mass chromatograms. In this paper, we report an improved version of the PRE algorithm that utilizes a second variable selection filter based on cross-correlation (CC). As a check on the ability of PRE and CC to select high quality mass chromatograms, every mass chromatogram in our data set (1451 in total) was individually inspected and rated as either high quality (green), intermediate quality (yellow), or poor quality (red). A color-coded plot of the CC value vs. the PRE value for the mass chromatograms was created, which shows that, as expected, the higher quality mass chromatograms are localized in its upper left quadrant, which corresponds to lower PRE values and higher CC values. In our original paper on this topic, we recommended a threshold of 0.5 sigma for PRE, which caused the algorithm to select 151 mass chromatograms out of 1451. Of these, 98 were of high quality, 6 were of intermediate quality, and 47 were of poor quality. Using a second threshold for CC, the algorithm retains all the high and intermediate quality mass chromatograms, while removing all 47 of the poor quality ones. The resulting TICC

关键词： Remote sensing   Gamma-ray spectroscopy   Cesium-137   Cobalt-60   Pattern recognition   Airborne detection  

摘要： A generalized methodology was developed for automating the detection of radioisotopes from gamma-ray spectra collected from an aircraft platform using sodium-iodide detectors. Employing data provided by the U.S Environmental Protection Agency Airborne Spectral Photometric Environmental Collection Technology (ASPECT) program, multivariate classification models based on nonparametric linear discriminant analysis were developed for application to spectra that were preprocessed through a combination of altitude-based scaling and digital filtering. Training sets of spectra for use in building classification models were assembled from a combination of background spectra collected in the field and synthesized spectra obtained by superimposing laboratory-collected spectra of target radioisotopes onto field backgrounds. This approach eliminated the need for field experimentation with radioactive sources for use in building classification models. Through a bi-Gaussian modeling procedure, the discriminant scores that served as the outputs from the classification models were related to associated confidence levels. This provided an easily interpreted result regarding the presence or absence of the signature of a specific radioisotope in each collected spectrum. Through the use of this approach, classifiers were built for cesium-137 (Cs-137) and cobalt-60 (Co-60), two radioisotopes that are of interest in airborne radiological monitoring applications. The optimized classifiers were tested with field data collected from a set of six geographically diverse sites, three of which contained either Cs-137, Co-60, or both. When the optimized classification models were applied, the overall percentages of correct classifications for spectra collected at these sites were 99.9 and 97.9% for the Co-60 and Cs-137 classifiers, respectively.

摘要： Feature extraction in pattern recognition has been researched for many years. Some works have been inspired by biological systems, particularly neural networks. As the biological system is not fully understood, this thesis proposed an algorithm inspired on the way humans see objects. Humans can tell two objects look similar when both objects have a high percentage of similarity. The similarity of individual features in both objects is measured before deciding how similar they are. Small features construct bigger features of the object, and at the end the combination of some bigger features make up the whole object. This process of recognising patterns following a hierarchical approach inspired the research presented in this thesis; hence the proposed algorithm is called the Hierarchical Concatenation Network (HCN). When two objects are compared, one object is used as the reference object in the comparison. Hence the HCN analyses the features of patterns using other patterns' features as reference. As features construct a whole object, the HCN proposed in this thesis has been designed to have several levels. The research investigated how many levels the HCN should have to produce better results. This led to investigate different approaches to concatenate low level features to represent the higher level features. Experimental results on measuring the similarity between specimens led the research to explore alternative ways to represent the high level features, hence this thesis reports on two feature activation methods called HCN-I and HCN-II. Preliminary results showed that the higher the level of the network, the more differentiated the two objects were. Exchanging the position of the tested and reference patterns produced different result. Hence, different methods of feature extraction were investigated to understand the effect of the relative position between the tested and reference patterns. This resulted in the proposed HCN having two methods called 'one to man

关键词： Internet of medical things   Deep deterministic learning   Electrocardiography   Cardiovascular diseases   Pattern recognition   Artificial neural network  

摘要： Electrocardiography (ECG) sensors play a vital role in the Internet of Medical Things, and these sensors help in monitoring the electrical activity of the heart. ECG signal analysis can improve human life in many ways, from diagnosing diseases among cardiac patients to managing the lifestyles of diabetic patients. Abnormalities in heart activities lead to different cardiac diseases and arrhythmia. However, some cardiac diseases, such as myocardial infarction (MI) and atrial fibrillation (Af), require special attention due to their direct impact on human life. The classification of flattened T wave cases of MI in ECG signals and how much of these cases are similar to ST-T changes in MI remain an open issue for researchers. This article presents a novel contribution to classify MI and Af. To this end, we propose a new approach called deep deterministic learning (DDL), which works by combining predefined heart activities with fused datasets. In this research, we used two datasets. The first dataset, Massachusetts Institute of Technology-Beth Israel Hospital, is publicly available, and we exclusively obtained the second dataset from the University of Malaya Medical Center, Kuala Lumpur Malaysia. We first initiated predefined activities on each individual dataset to recognize patterns between the ST-T change and flattened T wave cases and then used the data fusion approach to merge both datasets in a manner that delivers the most accurate pattern recognition results. The proposed DDL approach is a systematic stage-wise methodology that relies on accurate detection of R peaks in ECG signals, time domain features of ECG signals, and fine tune-up of artificial neural networks. The empirical evaluation shows high accuracy (i.e., 99.97%) in pattern matching ST-T changes and flattened T waves using the proposed DDL approach. The proposed pattern recognition approach is a significant contribution to the diagnosis of special cases of MI.

关键词： bridge   structural health monitoring   statistical pattern recognition   temperature effect  

摘要： This work describes a vibration-based structural health monitoring of a prestressed-concrete box girder bridge on the A100 Highway in Berlin by applying statistical pattern recognition technique to a huge amount of data continuously collected by an integrated monitoring system during the period from 2000 to 2013. Firstly, the general condition and potential damage of the bridge is described. Then, the dynamic properties are extracted from 20 velocity sensors. Environmental variability captured by five thermal transducers and traffic intensity approximately estimated by strain measurements are also reported. Nonlinear influences of temperature on natural frequencies are observed. Subsequently, the measurements during the first year are used to build a baseline health index. The multiple linear regression (MLR) method is used to characterize the nonlinear relationship between natural frequencies and temperatures. The Euclidean distance of the residual errors is calculated to build a statistical health index. Finally, the indices extracted from the following years gradually deviate;which may indicate structural deterioration due to loss of prestress in the prestressed tendons.