A learning framework for the optimization and automation of document binarization methods

Almost all binarization methods have a few parameters that require setting. However, they do not usually achieve their upper-bound performance unless the parameters are individually set and optimized for each input document image. In this work, a learning framework for the optimization of the binarization methods is introduced, which is designed to determine the optimal parameter values for a document image. The framework, which works with any binarization method, has a standard structure, and performs three main steps: (i) extracts features, (ii) estimates optimal parameters, and (iii) learns the relationship between features and optimal parameters. First, an approach is proposed to generate numerical feature vectors from 2D data. The statistics of various maps are extracted and then combined into a final feature vector, in a nonlinear way. The optimal behavior is learned using support vector regression (SVR). Although the framework works with any binarization method, two methods are considered as typical examples in this work: the grid-based Sauvola method, and Lu’s method, which placed first in the DIBCO’09 contest. The experiments are performed on the DIBCO’09 and H-DIBCO’10 datasets, and combinations of these datasets with promising results.     

An automatic diabetes diagnosis system based on LDA-Wavelet Support Vector Machine Classifier

In this paper, an automatic diagnosis system for diabetes on Linear Discriminant Analysis (LDA) and Morlet Wavelet Support Vector Machine Classifier: LDA–MWSVM is introduced. The structure of this automatic system based on LDA-MWSVM for the diagnosis of diabetes is composed of three stages: The feature extraction and feature reduction stage by using the Linear Discriminant Analysis (LDA) method and the classification stage by using Morlet Wavelet Support Vector Machine (MWSVM) classifier stage. The Linear Discriminant Analysis (LDA) is used to separate features variables between healthy and patient (diabetes) data in the first stage. The healthy and patient (diabetes) features obtained in the first stage are given to inputs of the MWSVM classifier in the second stage. Finally, in the third stage, the correct diagnosis performance of this automatic system based on LDA–MWSVM for the diagnosis of diabetes is calculated by using sensitivity and specificity analysis, classification accuracy, and confusion matrix, respectively. The classification accuracy of this system was obtained at about 89.74%.     

临床试验逻辑核查的分类及应用

临床试验数据的清理是临床试验过程中非常重要的环节,完整、清洁的数据才能用于统计分析。本文旨在讨论临床试验数据清理过程中数据核查计划的撰写及逻辑核查的分类等,并进一步讨论近年来临床试验数字化趋势对数据的采集及清理可能带来的影响。     

面向等离子体W材料改善韧性的方法与机制

目的: 探讨胰岛素样生长因子1（IGF-1）基因多态性与抗高血压药物治疗后血压水平的关联性。方法: 运用流行病学现况调查方法,收集服用复方利血平、降压片和珍菊降压片的780例高血压患者的一般临床特征并测量其血压和临床生化指标。选择IGF-1基因的5个标签位点,采用TaqMan技术进行基因分型。采用协方差分析及分层分析对其基因多态性与药物治疗后血压水平的关系进行评价。结果: 在服用降压片的高血压病人中,rs6219位点不同基因型之间舒张压水平差异有统计学意义（P=0.01）。在服用复方利血平的高血压病人中,rs5742612和rs6218位点不同基因型之间收缩压水平差异有统计学意义（P值分别为0.04、0.017）。在服用珍菊降压片的高血压患者中,rs6218位点不同基因型之间舒张压水平及rs6219位点不同基因型之间收缩压水平差异均有关联（P值分别为0.017、0.033）。进一步分层分析发现,服用降压片的患者中,随着rs35767位点遗传变异女性收缩压的水平呈线性下降,而男性舒张压水平呈线性上升,校正混杂因素后P值分别为0.028、0.038。结论: IGF-1基因多态性与复方利血平、降压片和珍菊降压片三种抗高血压药物治疗后血压水平均有显著相关,提示IGF-1基因可能是高血压重要的药物疗效生物标记。     

纯净水中病毒三种富集方法的比较

目的 比较膜吸附法、膜吸附-PEG法和膜吸附-超滤法对纯净水中病毒的富集效果,并对优选的富集方法的实验条件进行比对,旨在寻找高效的方法,为其他水体中病毒富集检测提供技术借鉴。方法 本文以MS2噬菌体为目标病毒,制作高、低浓度水样,建立逆转录-荧光定量聚合酶链式反应(RT-qPCR)方法的标准曲线,进行定量检测。使用膜吸附法、膜吸附-PEG法和膜吸附-超滤法富集水样,并比较不同过滤膜、不同超滤管及两种核酸提取方法的回收率,选择最适方法。结果 膜吸附-超滤法、膜吸附-PEG法和膜吸附法回收率均值分别为19.80%±12.19%、12.71%±9.09%和9.05%±4.89%,各因素对回收率的影响顺序为:膜吸附-超滤法>膜吸附-PEG法>膜吸附法;高浓度>低浓度。混合纤维素酯膜的富集效果优于尼龙膜。100 K和50 K的超滤管富集效果无差异。硅胶离心柱法和磁珠法核酸提取效果无差异。结论 通过二次富集可提高富集效率,膜吸附-超滤法优于另两种方法,且其操作简单、耗时短、对设备依赖性低,易于推广。膜吸附-超滤法的实验条件仍可优化、完善,拓展其在不同水体和不同病毒富集中的应用。     

梅花鹿茸血冻干粉的安全性评价

通过小鼠急性毒性试验,大鼠30 d喂养试验,三项遗传毒性试验,小鼠尾部铅皮负重游泳,小鼠血清中尿素水平检测,小鼠肝脏中糖原储备量测定以及小鼠运动前后血清中乳酸水平的测定对梅花鹿茸血冻干粉的食品安全性及对小鼠的缓解体力疲劳作用进行探讨。结果表明：小鼠急性经口毒性>15 g/kg·BW,属无毒级;鹿茸血冻干粉0.6,1.3,2.0 g/(kg·BW·d)剂量在30 d喂养试验中未对大鼠的体重、摄食量、血液常规和血液生化指标产生影响,且三项遗传毒性试验结果均为阴性。0.6 g/(kg·BW·d) 组小鼠负重游泳时间长于阴性对照组,差异有统计学意义（P<0.05）;0.1,0.2,0.6 g/(kg·BW·d) 组肝糖原储备量高于阴性对照组,差异有统计学意义（P<0.05）;0.6 g/(kg·BW·d)组小鼠游泳前后血乳酸变化的曲线下面积值低于阴性对照组,差异有统计学意义（P<0.05）。梅花鹿茸血冻干粉食用安全,且具有缓解小鼠体力疲劳功能。     

Thermal-Disrupting Interface Mitigates Intercellular Cohesion Loss for Accurate Topical Antibacterial Therapy

Bacterial infections remain a leading threat to global health because of the misuse of antibiotics and the rise in drug-resistant pathogens. Although several strategies such as photothermal therapy and magneto-thermal therapy can suppress bacterial infections, excessive heat often damages host cells and lengthens the healing time. Here, a localized thermal managing strategy, thermal-disrupting interface induced mitigation (TRIM), is reported, to minimize intercellular cohesion loss for accurate antibacterial therapy. The TRIM dressing film is composed of alternative microscale arrangement of heat-responsive hydrogel regions and mechanical support regions, which enables the surface microtopography to have a significant effect on disrupting bacterial colonization upon infrared irradiation. The regulation of the interfacial contact to the attached skin confines the produced heat and minimizes the risk of skin damage during thermoablation. Quantitative mechanobiology studies demonstrate the TRIM dressing film with a critical dimension for surface features plays a critical role in maintaining intercellular cohesion of the epidermis during photothermal therapy. Finally, endowing wound dressing with the TRIM effect via in vivo studies in S. aureus infected mice demonstrates a promising strategy for mitigating the side effects of photothermal therapy against a wide spectrum of bacterial infections, promoting future biointerface design for antibacterial therapy.     

Spark平台下的凸包问题研究

随着移动互联网时代的到来,越来越多的含地理位置信息的空间数据需要处理,如何在海量的空间数据中进行常见的几何查询成为一个挑战,凸包问题因其在模式识别、图像处理、统计学、地理信息系统、博弈论、图论等领域中被广泛应用成为近些年研究的一个热点。凸包问题的研究始于单机版的算法,进而过渡到Hadoop等基于硬盘的分布式系统,但是受限于单节点的计算存储能力的瓶颈以及Hadoop平台基于硬盘的特性,其计算性能尚不能达到人们的在线实时计算的需求。研究基于内存的分布式计算框架Spark下的凸包问题,给出基于Spark平台的凸包查询整体框架,框架从查询接口、语法解析和物理执行等多方面结合SparkSQL引擎。随后,给出基于Andrew单调链算法的单机算法CHStand,分析单机算法并行度上的问题后,提出基于Spark的CHSpark算法,进一步优化算法并提出一种Spark平台下的优化算法CHGeom。通过实验对比说明三种算法的相对性能提升,实验发现Spark平台下的解决方案相对传统的单机平台下的解决方案有着较大的性能提升,所提算法具有良好的拓展性和广泛的实际应用价值。