基于三次样条层间插值的多模态医学图像配准

我们从PET-CT多模态图像序列的特点出发,提出了一种全新的图像配准及融合方法,它采用三次样条插值法对PET-CT图像进行层间插值,然后再利用最大互信息法进行配准,最后应用改进的主成分分析(PCA)法融合PET-CT图像用以增强PET显像效果,从而得到满意的配准以及融合结果。用三次样条插值法进行层间插值并恢复层间缺失图像的信息,弥补了现有配准方法的不足,提高了配准精度,使融合后的图像更加接近实际的物理断层。该方法已经成功应用于三维适形放疗(3D-CRT)系统的开发中。     

Improved procedure of complex demodulation and an application to frequency analysis of sleep spindles in EEG

Complex demodulation is a local version of harmonic analysis that enables the amplitude and phase of particular frequency components of a time series to be described as functions of time. The paper presents a computational procedure involving complex demodulation with interpolation of data in the frequency domain. A computational procedure comprising repeated use of complex demodulation is also presented. This is used to estimate the optimum choice of the demodulating frequency which considerably influences the measurement of the instantaneous amplitude and phase of the underlying process. The usefulness of this procedure is verified by computer simulation. An example of applying this procedure to the estimation of the centre and the instantaneous frequencies of sleep spindles in the EEG (electroencephalogram) is presented. By using the procedure developed here, several partially overlapping sleep spindles are detected and correctly separated. The paper also presents an approach to separating and analysing transient time series (such as overlapping sleep spindles) by using an accurate frequency processing technique.     

发射计算机断层照相(ECT)的图像重建技术

本论文阐述的是断层重建技术在核医学领域的应用。通过投影滤波技术把病灶在二维空间的图像,扩展到三维空间,得到病人病灶纵深的“信息”,使临床医生更好、更快的定位病灶和了解病情。     

Three dimensional digital atlas of the orbit constructed from multi-modal radiological images

The human orbit has numerous structures packed in a relatively small space, the study of which is essential and difficult due to complex three dimensional relationships. Available printed orbital atlases do not convey the three dimensional information and are not interactive. To overcome these limitations, we built a digital 3D orbital atlas presented in axial, coronal and sagittal planes, and as three dimensional geometric models of the muscles, bones, and eyeball. The bone models are from a CT scan, the muscle and optic nerve from a MR scan, and other components that cannot be distinguished radiologically are modeled as geometric primitives from anatomic literature. All multi-modal data including the models and images are registered into the same space to form a complete atlas. All structures in the atlas are labeled with their names. An atlas browser is developed for user-friendly manipulation and presentation of the atlas content. Each structure can be turned on or off, rotated, zoomed, or moved, either individually or in unison with other selected structures. Thus, the relationships between different structures can be studied in greater depth. The method developed to build the orbital atlas is general and can be used to create other atlases or to build patient specific geometric models. The orbital atlas may be used for studying the orbital anatomy, as a reference guide for practitioners, and as a base for simulation of orbital surgery.     

Computer programs for calculation of median effective dose (LD50 or ED50) using the method of moving average interpolation

Over the past 40 years, toxicologists and pharmacologists have used tables published by Weil for the determination of LD₅₀ (or ED₅₀) values and their associated 95% confidence intervals. With the advances in computer technology, it is now common for investigators to have personal computers in their laboratories. Therefore, two identical programs were developed for determination of the LD₅₀ (or ED₅₀) which may be run on a personal computer. One of these programs was written in BASIC, and the other in FORTRAN. The programs are easy and rapid to use, requiring minimum computer hardware and little, if any, knowledge of programming. They also offer more user flexibility than the previously published tables of Weil, in that there are fewer restrictions on the number of animals and number of dosage levels used in an experiment. The output of the programs may be typed on the screen of a computer monitor, and may be sent to a printer. The two programs calculate the LD₅₀ and 95% confidence intervals for the LD₅₀. These programs should be valuable for many investigators.     

螺旋CT图像锥形伪影的模拟与实验

目的研究螺旋CT插值算法引起的锥形伪影的特征。方法根据插值算法和反投影重建基本原理建立重建图像的数学模型,阐明锥形伪影与扫描参数和被扫描物属性的函数关系;采用自制体模在SiemensPlus4扫描仪上进行实验验证。结果理论分析发现,伪影占图像的比重与扫描仪机架旋转1周的进床距离和锥形体模半锥角的正切值成正比,与重建位置的体模半径成反比。体模实验结果与理论分析相符。结论插值算法引起的锥形伪影与扫描参数和被扫描物属性存在确定关系。     

Ectopic Beats in Heart Rate Variability Analysis: Effects of Editing on Time and Frequency Domain Measures

Background: Various methods can be used to edit biological and technical artefacts in heart rate variability (HRV), but there is relatively little information on the effects of such editing methods on HRV. Methods: The effects of editing on HRV analysis were studied using R‐R interval data of 10 healthy subjects and 10 patients with a previous myocardial infarction (Ml). R‐R interval tachograms of verified sinus beats were analyzed from short‐term (～5 min) and long‐term (—24 hours) recordings by eliminating different amounts of real R‐R intervals. Three editing methods were applied to these segments: (1) interpolation of degree zero, (2) interpolation of degree one, and (3) deletion without replacement. Results: In time domain analysis of short‐term data, the standard deviation of normal‐to‐normal intervals (SDANN) was least affected by editing, and 30%‐50% of the data could be edited by all the three methods without a significant error (< 5%). In the frequency domain analysis, the method of editing resulted in remarkably different changes and errors for both the high‐frequency (HF) and the low‐frequency (LF) spectral components. The editing methods also yielded in different results in healthy subjects and AMI patients. In 24‐hour HRV analysis, up to 50% could be edited by all methods without an error larger than 5% in the analysis of the standard deviation of normal to normal intervals (SDNN). Both interpolation methods also performed well in the editing of the long‐term power spectral components for 24‐hour data, but with the deletion method, only 5% of the data could be edited without a significant error. Conclusions: The amount and type of editing R‐R interval data have remarkably different effects on various HRV indices. There is no universal method for editing ectopic beats that could be used in both the time‐domain and the frequency‐domain analysis of HRV. A.N.E. 2001;6(1):5–17     

Modelling the spatial distribution of Echinococcus multilocularis infection in foxes

Alveolar echinococcosis is a rare but fatal disease in humans and is caused by the fox tapeworm Echinococcus multilocularis. The densities of fox and grassland rodent populations and the interactions between them influence E. multilocularis transmission rates in Europe. Successful rabies control has caused fox populations and E. multilocularis prevalence rates to increase in many European countries. The potential increase of the infection pressure on the human population motivates the monitoring of the infection status of foxes over space and time. Detection of E. multilocularis antigen levels in fox faecal samples collected in the field might provide a pragmatic methodology for epidemiological surveillance of the infection status in wildlife hosts across large areas, as well as providing an indication of the spatial distribution of infected faeces contaminating the environment. In this paper, a spatial analysis of antigen levels detected in faeces collected in the Franche-Comté region of eastern France is presented. In Franche-Comté, rodent outbreaks have been observed to originate in areas rich in grassland. Spatial trends in fox infection levels were modelled here as a function of the composition ratio of grassland in the landscape derived from the CORINE land-cover map. Kriging models incorporating the grassland trend term were compared to a variety of models in which five alternative trend expressions were used: the alternative trend expressions included linear and quadratic polynomials on the x and y coordinates with and without a grassland term, and a constant mean model. Leave-one-out cross-validation indicated that the estimation errors of kriging with a trend models were significantly lower when the trend expression contained the grassland index term only. The relationship between observed and predicted antigen levels was strongest when the estimated range of autocorrelation was within the home range size of a single fox. The over-dispersion of E. multilocularis in foxes may therefore account for the majority of spatial autocorrelation locally, while regional trends can be successfully modelled as a function of habitat availability for intermediate hosts.     

桂林市道路交通伤害GIS反距离加权插值法分析

目的探讨广西桂林市2000—2009年道路交通伤害(RTI)死亡的地理分布和集中趋势,为预防和减少RTI的发生提供科学依据。方法以桂林市电子地图为背景,利用ArcGIS 9.2地统计分析模块中的反距离加权(IDW)插值法绘制桂林市RTI死亡的分布地图。结果 2000—2009年桂林市秀峰区、叠彩区、象山区、七星区和雁山区发生RTI合计3 603次,死亡506例,受伤2 911例,直接经济损失为1 011.2万元;2000—2008年每年发生RTI次数均以象山区最多,2009年以七星区最多;2000—2004年桂林市RTI造成的死亡人数以雁山区和象山区最多;2005—2006年以叠彩区、象山区和雁山区较多;2007—2009年以七星区、象山区和雁山区较多;2000—2009年桂林市RTI死亡人数分布的IDW插值结果表明,RTI导致的死亡人数以雁山区最多,其次为象山区。结论 2000—2009年桂林市RTI导致的死亡人数以雁山区最多;IDW插值法制图结果较为可信。