基于切面识别的房间隔缺损智能辅助诊断

针对超声心动图像质量差、噪声多,传统卷积神经网络架构对超声心动图像的学习能力有限、表达不充分的缺点,提出了一种基于标准切面识别的房间隔缺损（Atrial septal defect,ASD）智能辅助诊断模型。该模型通过对超声心动图像进行切面识别,充分融合其不同切面的语义特征,使得诊断的准确率得到明显提升。此外,还对其进行双边滤波保边去噪,并基于此模型搭建房间隔缺损智能辅助诊断系统（简称ASD辅助诊断系统）。结果表明,该ASD辅助诊断系统的准确率高达97.8%,且与传统卷积神经网络相比大大降低了假阴性率。      

基于多源遥感数据融合的矿区土地利用分类方法研究

本文旨在探究利用多源遥感数据融合结合机器学习算法实现土地利用分类的可行性。以冬瓜山矿区及其周边区域作为研究对象,基于NND(Nearest Neighbor Diffusion)融合算法实现哨兵2(Sentinel-2)数据与中巴地球资源卫星04星（CB04）全色数据的融合,并基于融合后的5m分辨率影像实现了建模方法的优选,进行了全区范围内的土地利用分类。实验表明,利用影像融合方法可在保留哨兵2光谱信息的同时有效提升空间分辨率。而在三种机器学习算法下,基于随机森林算法所建分类器整体精度较高,其Kappa系数达到了0.9以上。上述结果则表明,利用多源遥感数据融合与随机森林算法来实现土地利用分类是可行的,该方法不仅可以有效提升影像的空间分辨率,也在一定程度上实现了土地利用分类细化。     

基于参考模型的视网膜特征量化

提出一种基于参考模型的视网膜特征量化方法，结合医生诊断过程中关注的视网膜形态变化特征，提出一系列适用于计算机判断分析视网膜状态的可量化特征.在完成正常光学相干断层成像（OCT）中视网膜内界膜（ILM）、光感受器内外节交界处（ISOS）、布鲁赫膜（BM）分割提取的基础上，利用统计方法构建正常视网膜参考模型.结合参考模型和医生所关注的视网膜厚度、边界平滑度以及边界连续性，实现视网膜不同区域厚度特征、厚度比值特征、梯度特征、曲率、标准差、相关系数特征的计算.基于正常OCT图像所构建的参考模型，获取了正常视网膜的厚度及形态特征量化数值.通过分析比较异常OCT图像与参考模型特征数值之间的差异，可以对应表征出异常图像中病变导致的异常形态所在位置及严重程度.实验结果表明，通过参考模型获得的正常视网膜特征信息可以为医生提供数值参考，同时对异常OCT图像量化得到的特征数值可以表现出图像中的异常形态，为后续的异常判断提供基础.      

基于颜色特征的茄子病害图像检索方法

为克服当前蔬莱病害诊断专家系统依靠文字提供信息的不准确性,提出一种基于颜色特征的茄子病害图像检索方法.详细分析茄子病害图像的颜色特征,在符合人眼视觉特征的HIS颜色空间上进行图像预处理.提取图像的均值、方差、偏度、峰度、能量作为检索特征值,构建图像特征矩阵.利用高斯归一化方法进行归一化处理后,采用欧式距离进行相似性度量.在Visual C+ +6.0开发环境下,采用C++编程开发了基于颜色特征的茄子病图像检索系统.结果表明,基于颜色特征的茄子病害图像检索方法的查准率为65%,查全率为83%,识别效果较好.将该算法应用于茄子病害诊断专家系统,将大大提高系统的鲁棒性,能够满足病害诊断的要求.     

基于多目标支持向量机的ADHD分类

注意力缺陷多动障碍（ADHD）是儿童期最常见的精神疾病之一,在大多数情况下持续到成年期。近年来,基于功能磁共振数据的ADHD分类成为了研究热点。文献中已有的大多数分类算法均假设样本是均衡的,然而事实上,ADHD数据集通常是不平衡的。传统的学习算法会使得分类器倾向于多数类样本,从而导致性能下降。本文研究了基于不平衡神经影像数据的ADHD分类问题,即基于静息状态功能磁共振数据对ADHD进行分类。采用功能连接矩阵作为分类特征,提出了一种基于多目标支持向量机的ADHD数据分类方案。该方案将不均衡数据分类问题建模为具有三个目标的支持向量机模型,其中三个目标分别为最大化分类间隔、最小化正样本误差和最小化负样本误差,进而正负样本经验误差可以被分开处理。然后采用多目标优化的法向量边界交叉法对模型进行求解,并给出一组代表性的分类器供决策者进行选择。该方案在ADHD-200竞赛的五个数据集上进行测试评估,并与传统分类方法进行对比。实验结果表明本文提出的三个目标支持向量机分类方案比传统的分类方法效果好,可以有效的从算法层面解决数据不平衡问题。该方案不仅可用于辅助ADHD诊断,还可用于阿尔茨海默病和自闭症等疾病的辅助诊断。      

基于FCM-WA联合算法的多种类矿石图像分割

矿石图像分割是基于机器视觉的矿石粒度分布检测的重要组成部分。针对复合矿山中颜色多样、纹理复杂且边缘粘连的多种类矿石图像难以识别与分割的问题,提出了一种基于FCM-WA联合算法的矿石图像分割方法。首先对矿石图像进行形态学优化,利用双边滤波、直方图均衡化和形态学重构来优化矿石图像的几何特征,减少噪声对分割效果的影响,提高图像对比度;然后将模糊C均值聚类（FCM）算法与分水岭（WA）算法相结合,利用FCM算法进行聚类迭代,计算出合适的分割阈值并对矿石图像进行分割,输出二值化图像;再利用基于距离变换的WA算法优化FCM算法的分割结果,对FCM算法输出的矿石图像边缘粘连部分进行分割,以获取最佳的分割图像。研究结果表明：（1）利用形态学优化流程处理矿石图像能够减少噪声并增强边缘信息,从而提高对比度;（2）相比传统的大津法和遗传算法,本文所提FCM-WA方法的稳健性更强、分割效果更好,对多种类的矿石图像像素分割准确率和矿石粒度识别准确率均可达到92%以上;（3）通过试验验证,FCM-WA方法能够精确地分割颜色多样、纹理特征复杂及边缘粘连的多种类矿石图像,分割结果满足粒度分布检测的要求;（4）FCM-...     

电工钢在线铁损数据分析及钢卷评价系统建立

电工钢生产线上都具有在线铁损曲线数据,为现场产品质量的及时诊断提供了有价值的信息,但不少企业对该数据尚未充分利用,既未将其用于产品整卷质量定量评价,也未用于考察工艺参数对磁性影响的统计性分析和各种质量问题的自动诊断。本文结合对现场数据的一些了解,以及检测现场工艺参数调整时磁性能变化遇到的困难,提出利用这些数据定量整体评价整卷质量的思路,并初步编制了相关计算机软件,随着该系统在实践运用中的进一步完善,必将可以更好利用这些数据精确评价、控制整卷产品的质量。     

基于无人机遥感测量的铝土矿测绘研究

提出了一种利用无人机搭载遥感设备,利用卫星影像数据的光谱分解技术实现对XX区域铝土矿富集地区的矿物含量的定位和量化,业界已将ASTER影像(VNIR和SWIR波段)与SRTMDEM结合使用,并在野外探测领域得到了验证。利用ASTER图像数据光谱分解的新方法勾勒出富含氧化铝的区域,同时可以获取该地域铝土矿新形成的各种地质与地貌参数;利用各种地质与地貌参数可准确的分析出铝土矿的地域位置,而且还可估算矿床的质量。     

地质找矿预测中三维地形与二维影像数据集成研究

矿产资源开发对于促进国民经济发展有着重要的作用,而地质找矿预测的精度和准确性直接影响到矿产开发工作。基于此,为了辅助地质找矿预测工作,结合实例分析,提出了一种三维地形与二维影像多源数据集成方法,以GIS软件为基础,建立地形的三维影响模型,再借助二维影像方法,结合三维模型与多源数据集成,进行地质解释,这种方法能够有效提升地质找矿预测的准确性、科学性和勘查精度,对后期矿产资源的开发工作有着重要的指导作用。     

信息技术——英语教学的调色盘

利用多媒体、网络资源辅助教学已成为21世纪教学的主题.利用多媒体、网络资源应用于高中英语教学,可以使教学活动集文字、声音、图像、影像、动画等信息功能于一体,能最大限度地调动、激发学生的学习积极性和主动性,实现课堂教学的大容量、多信息、多趣味和高效率.     

Discrepancies between patient recall and the medical record. Potential impact on diagnosis and clinical assessment of chronic disease

BACKGROUND: During a case-control study, data necessary for fulfilling diagnostic and classification criteria for spondyloarthropathy were collected from 121 patients. OBJECTIVE: To study the potential impact of differences between patient recall and the medical record on diagnosis and clinical characterization of spondyloarthropathy as a model of chronic disease. METHODS: The study was conducted among four Alaskan Eskimo populations served by the Alaska Native Health Service. Two sets of historical data were compiled for each subject, one acquired during the interview and the other derived from the medical record. Paired items from the interview and the medical record were analyzed to determine discrepancies and consequent effects on diagnosis, classification, and disease characterization. RESULTS: Significant differences were observed in the reporting of genitourinary or diarrheal illnesses preceding or associated with arthritis, the occurrence of eye inflammation in association with joint pain, the occurrence of joint pain and back pain together, and the age at onset of back pain all of which are important to the diagnosis and classification of spondyloarthropathy. In contrast, for information needed to establish the probable inflammatory nature of back pain, patient interview was more helpful than the medical records, which did not provide adequate details to differentiate inflammatory from mechanical back pain. CONCLUSIONS: Patient recall bias can substantially affect diagnosis and clinical assessment of chronic disease, as exemplified by spondyloarthropathy. Reliance on records alone, however, may lead to underestimation of features that require subjective appraisal by the patient.     

基于脑电多视图混合神经网络的时空半监督睡眠分期

睡眠分期是评价睡眠质量的必要基础,现阶段的工作大部分采用全监督学习和单一维度视图信息进行,这不仅需要技师进行大量的睡眠数据标注,还可能因特征提取不充分而导致分期准确率受限的问题。利用半监督学习策略,实现对脑电无标注数据的学习。提出一种多视图混合神经网络,首先用多通道视图时频域机制分别提取时域信号特征和空域信号特征,实现多视图特征提取;再通过注意力机制加强对显著性特征的提取;最后将上述混合特征融合并分类。在三个公开数据集和一个私有数据集中与全监督学习进行了对比评估,半监督学习取得平均准确率为81.0%,卡帕值为73.2%。结果表明,本文模型可以与全监督学习的睡眠分期模型相媲美,同时显著减少技师标注数据的工作量。      

Psychopathological assessment and diagnosis. A study of specificity of single symptoms

The importance of single symptoms in the diagnostic process has not been adequately studied, mainly because of the complexity of the information involved. The aim of the present study is to investigate the interconnection between psychopathological assessment and diagnostic classification, with the aid of a nonparametric, inductive pattern recognition method. Using the concept of inductive logic and a theory of inductive knowledge acquisition, a model has been developed to describe psychopathological assessment and diagnostic classification. Based on a comprehensive psychopathological assessment in 837 patients with 14 different diagnoses, classification values were calculated for specific symptoms in different diagnostic groups. Furthermore, nonparametric statistical procedures have advantages over discriminant analytic approaches: more information is utilized in differentiating the groups and differentiations can be made between more groups, whereby the rate of correctly classified cases is comparable with discriminant analytic approaches. The pattern recognition method appears to illustrate the multidimensional, medical decision-making in a comprehensible way.     

基于深度学习的辊道电机故障诊断分析方法

为了避免电机频繁无预警停机,机电设备诊断分析是一项非常重要的任务。针对辊道电机中单个电机故障,根据所有电机正常运行状况和故障状况下收集的电机电流数据,提出一种基于深度学习的智能化电机故障诊断分析策略。首次提出了考虑用协同工作的其他电机的电流来对被监测电机进行故障诊断。该方案选择采用Lenet-5模型进行分类预测训练,以诊断电机中的故障,在生产现场实际应用此方案并进行数据结果验证。结果表明,所提出的方法在电机故障诊断中可行且有效。     

The clinical diagnosis of deep venous thrombosis: integrating incidence, risk factors, and symptoms and signs

Symptoms and clinical signs suggestive of deep venous thrombosis (DVT) are common and have numerous possible causes. Studies have shown that symptoms and clinical signs in themselves are inaccurate for the diagnosis of DVT. However, clinicians have other information at hand, such as data on risk factors for DVT, that may help improve their ability to predict a diagnosis of DVT in the individual patient. Epidemiological data on DVT incidence and risk factors were reviewed, as were published data on the accuracy of clinical diagnosis of DVT, with the use of both symptoms and signs in isolation and symptoms and signs combined with other clinical information in the form of clinical prediction indexes. Symptoms and clinical signs, when combined with other patient information such as the presence or absence of known risk factors for DVT, can improve clinical prediction considerably. Further study is needed to determine whether clinical prediction indexes have a role in improving the diagnostic process in patients with suspected DVT.     

GENFILES: a computerized medical genetics information network. II. MEDGEN: the clinical genetics system

MEDGEN, a clinical genetics information storage and retrieval system, facilitates the handling of medical records for the central genetics clinic and satellite clinics conducted by the University of California, San Francisco. The system is part of the GENFILES genetics network, which handles all of the genetics data generated by a comprehensive medical genetics center. The clinical data stored on each patient include 1) diagnoses, which utilize McKusick catalog numbers as well as our own diagnostic codes; 2) relevant medical, gestational, and pregnancy history; 3) clinical manifestations (functional and structural); 4) karyotype information through a crosslink to the cytogenetics file; 5) ethnic origin of the patients; 6) physical status and sex of the patient; 7) laboratory studies, including results of metabolic tests; and 8) any additional remarks deemed necessary for complete understanding. The data, staff member attending, and physical location of each visit also are recorded.     

Clinical approach to the diagnostic evaluation of progressive neuromuscular diseases

This article has reviewed the clinical approach to the diagnostic evaluation of progressive neuromuscular diseases with an emphasis on relevant neuromuscular history, family history, clinical examination findings, laboratory studies, and a brief discussion of the role of muscle biopsy. Molecular genetic and immunocytochemistry studies of muscle have been major advances in the diagnostic evaluation of the neuromuscular disease patient; however, all diagnostic information must be interpreted within the context of relevant clinical information. In some instances, a precise diagnosis is not medically possible; however, the accurate characterization of an individual patient within the most appropriate NMD clinical syndrome often allows the clinician to provide the patient and family with accurate prognostic information and anticipatory guidance for the future. After synthesizing all available clinical and diagnostic information, the physiatrist or neurologist may at times determine that an NMD patient has an inappropriate diagnosis warranting further diagnostic evaluation. This issue focuses on the rehabilitation of progressive neuromuscular diseases with an emphasis on optimization of health, prevention or minimization of complications, and enhancement of quality of life. Appropriate rehabilitation approaches require an accurate diagnosis. In addition, patient quality of life in NMD depends on access to current and accurate information. The first step in providing accurate information and appropriate treatment is constantly ensuring that NMD patients have appropriate diagnoses based on a through evaluation of clinical information and appropriate application of current medical science and available diagnostic technology.     

基于空洞卷积神经网络与注意力机制GRU的滚动轴承故障诊断

针对传统故障诊断方法需要人工提取特征的不足,以及大数据下滚动轴承故障振动信号自适应特征提取与智能诊断问题,利用空洞卷积神经网络(DCNN)可以在不增加计算量的基础上兼顾不同尺度空间特征的能力、门控循环单元(GRU)善于从动态变化的序列数据中学习到时间上的关联性的能力,提出了一种将DCNN、注意力机制和GRU多路径融合的端到端故障诊断方法。首先利用DCNN从原始数据中自动提取时序信号特征,然后将注意力机制(Attention)的GRU通路和DCNN通路进行融合,最后将提取到的特征融合之后送入分类层进行分类。试验结果表明,所提方法的诊断准确率平均为98.75%,高于比较方法,更加适用于滚动轴承故障诊断。     

基于云理论的隧道结构健康诊断方法

隧道健康状态诊断过程作为一个模糊系统,具有随机性和模糊性的特点.传统的隧道结构健康诊断方法无法妥善地将系统的模糊性和随机性关联在一起.基于此,针对隧道结构健康状态诊断系统的模糊性和随机性特点,提出了基于云理论的隧道结构健康状态诊断方法.首先,建立反映隧道结构健康状况等级评语集的正态云模型.然后,依据逆向云发生器原理,将隧道结构健康状态指标的监测数据的归一化结果值转化为隶属度云模型.将健康状态指标的重要性语言值转化为权重云,用于表征各健康状态指标的重要程度.最后,运用云理论的计算方法对隧道结构健康状态等级进行诊断,得到健康状态等级诊断结果的云模型.应用云模型的改进方法,对某市地铁2号线盾构隧道结构健康状态进行了诊断,极大提高了诊断结果的可视化和鲁棒性.      

An exploration of clinicians' diagnostic approaches to premenstrual symptomatology.

Explored whether non-medical clinicians and medical clinicians use a medical and/or sociocultural model to diagnose women presenting with PMS. Equal groups of female and male, Anglophone and Francophone, medical and non-medical clinicians (aged 24–82 yrs) were chosen in random samples of Quebec's professional orders. All clinicians were asked to diagnose a female patient described in a scenario as having symptoms construed to look like PMS. While 73% of the clinicians gave a PMS diagnosis, only 5% indicated that they would use prospective data to confirm their diagnosis. Clinicians who gave a PMS diagnosis explored more medical issues in the intake interview than those who did not. Medical clinicians considered more medical issues and a larger proportion of medical issues than non-medical clinicians did. Anglophone clinicians, female clinicians, and clinicians who did not diagnose PMS considered a greater number of sociocultural factors than did Francophones, males, and clinicians who did diagnose PMS. (PsycINFO Database Record (c) 2010 APA, all rights reserved)