改进的基于核函数的FCM图像分割算法

目的：对含有噪声的脑部MR图像进行分割。方法：改进了传统FCM目标函数,引入核函数的概念,用内核诱导距离来代替传统的欧氏距离;考虑到邻近像素的影响,增加了空间约束项,从而提出了一种能够合理应用图像空间信息的、改进的基于核函数的模糊C均值聚类算法。结果：对叠加椒盐噪声模拟图分割,该算法分割结果显示各灰度区域内无噪声污点;对仿真脑部MR图像分割,该算法错分率低于基于核函数的FCM算法其及改进算法。结论：与传统的FCM和其他改进算法相比,该算法能够对非精确图像进行更精确的分割。     

基于模糊聚类空间模型的非均匀MR图像分割

模糊C-均值聚类算法（FCM）已广泛地运用到MR图像的分割中。但传统的算法中未利用图像的空间信息，在分割叠加了噪声和附加了偏移场的非均匀MR图像时分割效果不理想。本文改进了传统FCM的目标函数，引入控制邻域作用紧密程度的参数，提出了一种能够更加合理地运用图像的空间信息，改进的模糊C-均值聚类算法。通过对脑部MR图像的分割实验表明，该算法能够比传统的FCM和其它改进算法对非均匀图像进行更精确的分割。     

改进的模糊C均值聚类算法

由于传统的模糊C均值(fuzzy C-means,FCM)算法没有考虑像素点的空间邻域信息,仅涉及像素的单点灰度,在处理含有噪声的图像时有很大的局限性,因此分割效果较差.针对FCM的缺陷,提出一种新的改进算法,该算法引入Gibbs随机场,将Gibbs随机场先验概率与像素点隶属度的乘积作为新的像素隶属度.实验表明,改进后的算法有良好的分割效果,既可以较为完整地保留图像边界细节,又能较好地去除图像的噪声.     

Applicability analysis of five segmentation algorithms for high resolution remote sensing Images

本文将5种图像分割算法应用在高分辨率遥感图像分割上,并利用图像分割评价指标,对5种分割算法进行了对比分析,评价了各种方法的优缺点,讨论了它们在高分辨率遥感图像分割中的适用性,明确了不同分割方法的适用条件.实验结果表明,改进的分水岭分割法与JSEG分割法在高分辨率遥感图像分割中的适用性比较强,对大小斑块分割结果都比较好,而其他3种方法不能兼顾不同等级的斑块.     

基于改进区域生长算法的肝脏管道图像分割方法

在图像分割过程中,传统区域生长法种子点的选取需人工判定,工作量较大,效率较低。为了减少种子点选取时的用户交互量,本文提出了一种基于种子点位置预判的改进区域生长算法。该算法基于血管骨架线具有代表性的特点,通过坐标系的映射转换来预判图像中肝脏管道的位置,减少了人工参与,实现了图像种子点数目和位置的自动确定。肝脏图像分割实验结果表明,该算法在较少的用户交互情况下,实现了序列图像的种子点位置预判,获得了较满意的图像分割效果。     

基于纹理特征和GRBF网络的医学图像分割

目的：研究一种新的图像分割算法，以获得更好的医学图像的分割结果。方法：采用基于纹理特征和广义径向基函数（Generalized Radial Basis Function，GRBF）神经网络的图像分割方法对医学图像进行分割．根据灰度共生矩阵获得纹理特征参数，形成广义径向基函数神经网络的输入矢量，对网络进行训练和仿真测试。结果：该算法取得更为理想的图像二值化的结果，且实验发现GRBF网络的泛化能力比RBF网络更胜一筹。结论：研究了边界模糊的医学图像的纹理特征．结合GRBF神经网络一得到一种有效的医荤图像奔割方法     

基于内容的医学图像检索中对Tamura纹理特征的算法改进

目的：对Tamura纹理特征的算法进行改进,使之在医学图像检索中适用于不规则图形的计算。方法：对乳腺图像进行分割,提取出感兴趣区域（ROI）。在样本（包含正常乳腺图片和乳腺肿瘤图片共281幅）中使用传统算法Tamura纹理特征和改进算法的Tamura纹理特征进行图像内容检索,比较2种算法的检索结果。结果：初次检索时,传统Tamura算法在样本中检索出93幅相关图像,175幅不相关图像：改进后的算法检索出91幅相关图像．131幅不相关图像。结论：使用改进算法的Tamura纹理特征进行查询,在保证查全率的基础上查准率有较大提高。提示改进算法后的Tamura纹理特征更适用于医学图像中不规则区域的特征计算。     

基于改进测地线模型的前列腺超声图像分割

利用变分法,消除了测地线中的重初始化过程,使运算速度加快;结合先验知识,改进了测地线活动轮廓的边缘检测函数;使用最小方差,提高了分割算法的精度。实验表明,该方法能够有效地进行分割前列腺超声图像,并且避免边界泄露问题。     

基于改进C-V模型的乳腺肿瘤超声图像分割

提出了一种改进的C-V模型,完全避免了重初始化步骤并简化了初始水平集函数的构造,大大加快了分割速度。针对乳腺肿瘤超声图像灰度分布的特点和C-V模型分段常量的假设,提出了手工勾画粗略边界,再划分子图进行分割的半自动分割流程,不仅提高了分割准确性,同时也进一步提高了分割效率。实验表明,算法能高效准确地从超声乳腺肿瘤图像中提取出肿瘤的边界,为下一步的目标特征提取、分析打下了良好的基础。     

基于多尺度匹配滤波法的视网膜血管分割

视网膜血管功能复杂、形态结构因人而异,易受自身血管性疾病和全身血管性疾病的影响,其准确分割对疾病诊断、身份识别有重大意义。该研究针对视网膜血管尺度不均等问题创新性地提出了一种多尺度匹配滤波算法,在传统的单一尺度的高斯匹配滤波基础上,分别采用大小两个尺度的高斯匹配滤波器增强灰度图像,再使用二维最大熵阈值分割算法对叠加后图像进行二值化处理。该算法在DRIVE数据库中测试,灵敏度、特异性、准确度分别为0.803、0.959、0.981。与传统算法比,该算法灵敏度高、运行速度快、分割结果细节丰富。     

基于小波变换的多分辨率CT/MRI图像配准

本文以仿射模型为基础将区域配准问题转化成优化问题,然后给出小波金字塔分解用于配准的一般框架,并与常规金字塔分解法进行了分析比较;针对配准过程的多变量非线性优化问题,结合混沌理论和遗传算法提出了改进的全局优化的混沌遗传算法。这种算法将混沌序列嵌入标准遗传算法,对标准遗传算法的个体进行混沌化处理,利用混沌的遍历特性来改进遗传算法的鲁棒性。     

克隆算法在视网膜血管边缘的检测中的应用研究

在分析图像边缘克隆原理及算法的基础上,同边缘连接和噪声去除算法相结合,提出了一种边缘克隆算法。仿真实验结果表明,该算法检测出的边缘图像解决了传统边缘检测算法边缘不连续和噪声大的问题,具有更好的边缘识别性能。     

Evaluation of three algorithms to identify incident breast cancer in Medicare claims data

OBJECTIVE: To test the validity of three published algorithms designed to identify incident breast cancer cases using recent inpatient, outpatient, and physician insurance claims data. DATA: The Surveillance, Epidemiology, and End Results (SEER) registry data linked with Medicare physician, hospital, and outpatient claims data for breast cancer cases diagnosed from 1995 to 1998 and a 5 percent control sample of Medicare beneficiaries in SEER areas. STUDY DESIGN: We evaluate the sensitivity and specificity of three algorithms applied to new data compared with original reported results. Algorithms use health insurance diagnosis and procedure claims codes to classify breast cancer cases, with SEER as the reference standard. We compare algorithms by age, stage, race, and SEER region, and explore via logistic regression whether adding demographic variables improves algorithm performance. PRINCIPAL FINDINGS: The sensitivity of two of three algorithms is significantly lower when applied to newer data, compared with sensitivity calculated during algorithm development (59 and 77.4 percent versus 90 and 80.2 percent, p<.00001). Sensitivity decreases as age increases, and false negative rates are higher for cases with in situ, metastatic, and unknown stage disease compared with localized or regional breast cancer. Substantial variation also exists by SEER registry. There was potential for improvement in algorithm performance when adding age, region, and race to an indicator variable for whether the algorithm determined a subject to be a breast cancer case (p<.00001). CONCLUSIONS: Differential sensitivity of the algorithms by SEER region and age likely reflects variation in practice patterns, because the algorithms rely on administrative procedure codes. Depending on the algorithm, 3-5 percent of subjects overall are misclassified in 1998. Misclassification disproportionately affects older women and those diagnosed with in situ, metastatic, or unknown-stage disease. Algorithms should be applied cautiously to insurance claims databases to assess health care utilization outside SEER-Medicare populations because of uneven misclassification of subgroups that may be understudied already.     

一种用于自动体外除颤器的可电击复律心律检测算法

本文介绍了一种标准化斜率绝对值标准差算法,用于判断ECG中的心室纤颤（VF）和心动过速（VT）。使用被广泛认可的CU Database和AHA Database库对算法的灵敏性、特异性、正向预测率、正确率及计算时间等指标进行了测试,并与其它两种算法进行了对比,结果显示该算法具有很好的分类性能和更短的计算时间。     

基于降噪风险基因网络的生存风险基因筛选

目的 对基于降噪风险基因网络的生存风险基因筛选算法CoxLASSO-ISIS-N的表现作出评价,并与其他5种算法进行比较,分析算法的优劣,以期为高维基因表达数据的生存风险基因筛选提供一种新的思路.方法 针对带有噪声的高维基因表达数据的生存风险基因筛选问题,分别利用模拟数据和真实数据,对6种算法(CoxLASSO、Cox...     

Algorithms for verbal autopsies: a validation study in Kenyan children.

The verbal autopsy (VA) questionnaire is a widely used method for collecting information on cause-specific mortality where the medical certification of deaths in childhood is incomplete. This paper discusses review by physicians and expert algorithms as approaches to ascribing cause of deaths from the VA questionnaire and proposes an alternative, data-derived approach. In this validation study, the relatives of 295 children who had died in hospital were interviewed using a VA questionnaire. The children were assigned causes of death using data-derived algorithms obtained under logistic regression and using expert algorithms. For most causes of death, the data-derived algorithms and expert algorithms yielded similar levels of diagnostic accuracy. However, a data-derived algorithm for malaria gave a sensitivity of 71% (95% Cl: 58-84%), which was significantly higher than the sensitivity of 47% obtained under an expert algorithm. The need for exploring this and other ways in which the VA technique can be improved are discussed. The implications of less-than-perfect sensitivity and specificity are explored using numerical examples. Misclassification bias should be taken into consideration when planning and evaluating epidemiological studies.     

What is the use? Application of the short form (SF) questionnaires for the evaluation of treatment effects

Background

Health status has evolved as a clinical outcome measure that is of great interest in medical care. However, there is still debate about the appropriateness of scoring algorithms for the often used short form questionnaires. Therefore, our aim was to evaluate the consequences of the traditional scoring procedure based on orthogonal factor rotation for clinical applications by (a) re-evaluating the results of randomized controlled trials (RCTs) on the effectiveness of antidepressants in improving health status in cardiac patients and (b) comparing empirical evidence on depression and health status using orthogonal and oblique factor rotation (alternative scoring method) in a community sample and a heart failure (HF) sample.

Methods

This is a systematic literature review and cross-sectional analysis among 1,598 community sample participants and 282 HF patients.

Results

Orthogonal rotation artificially forces the mental component summary (MCS) and physical component summary (PCS) to be unrelated, which is illustrated in two of the three included RCTs. Two RCTs showed improvements in MCS, but no improvement in PCS over time. Cross-sectional analysis in the two datasets showed that employing the alternative scoring algorithm resulted in higher negative correlations of MCS and PCS with depression, and a gradual decline in MCS with each decile of decline in PCS.

Conclusion

Our data showed that appropriate carefulness is needed when calculating and interpreting summary scores. The traditional scoring algorithm seems inappropriate to objectively evaluate the effects of interventions on both the MCS and the PCS. Awareness in the design and evaluation of interventions using these outcomes is warranted.     

自组织映射模型在医学领域数据挖掘中的应用

目的 介绍自组织映射模型的基本原理、常见算法及其研究进展,阐述自组织映射模型在医学领域数据挖掘中的应用,以期为相关研究提供参考方法 本文从医学领域的数据挖掘角度出发,采用文献回顾的方法,阐述自组织映射模型在处理非线性、高度冗余数据时运算速度和精度上的优势,并分析自组织映射模型各种算法的特点及不足结果 相对于传统的数据挖掘方法在大规模的医学数据挖掘时无法消除数据冗余及捕捉非线性特征,从而导致挖掘过程速度慢、精度低等问题,自组织映射模型可以自动地、智能地、可视化地将待处理的数据转化为有用的知识,较好地弥补传统数据挖掘方法的不足.自组织映射模型算法的不断改进,保证了它可以适用于具有不同特征的数据结论 自组织映射模型在医学领域的数据挖掘中得到了较好的应用,其算法还有进一步改进的空间,在医学领域的应用前景将会得到进一步扩展.