一种快速有效的医学图像灰度插值算法

介绍一种用于快速有效的医学图像灰度插值算法。通过在两幅已知图像上寻找对应点，有效地解决了线性插值边界模糊的问题，同时避免了基于形状的插值算法中图像分割的问题，具有较强的适用范围。实验结果表明此算法能得到令人满意的插值结果。     

Evidence for curvilinear interpolation from dot alignment judgements

Visual interpolation between dots responsible for rectilinear versus curvilinear contour interpretation was examined with the psychophysical forced directional response (FDR) paradigm. Regular four-dot polygon segments, together with a target dot, were presented to the subjects for 150 ms. Subjects were required to indicate the direction of deviation of the target dot from the midpoint of the intermediate line segment. Crucial variables were the outer angle of the line segments and symmetry axis orientation of the polygon segment. Logistic regression analyses showed that curvilinear interpolation occurred for angles up to 30 degrees, but emerged more pervasively under the vertical symmetry axis orientation for angles up to 60 degrees.     

空间插值法在人群包虫病患病率预测中的应用

目的探索不同空间插值方法在人群包虫病患病率预测中的应用并比较其效果。方法以四川、西藏、甘肃和青海四省(区)青藏高原地区2012年全国包虫病流行情况调查获得的124个藏区县的人群包虫病患病率数据为基础,使用Arcgis10.0软件,采用局部多项式插值法、反距离权重插值法和kriging插值法进行空间插值,比较青海省治多县和玛多县人群包虫病患病率推算的效果。结果 3种方法预测的精度各有优势,对青海省治多县和玛多县人群包虫病患病率误差最小的为Kriging插值方法,预测的患病率分别为2.04%和5.28%。结论空间插值方法可用于青海省治多县和玛多县人群包虫病患病率预测,以Kriging插值法的预测效果较好。     

Blind method for low-order interpolation detection in digital images

Digital images are easily forged due to the availability of powerful imaging hardware and image editing software. Copy-move and splicing are the most common type of image forgeries and various post-processing operations including geometric transformation (resizing and rotation) are employed by the attacker to conceal the visual traces of doctoring operations. In this brief paper, a blind algorithm is proposed to detect global rescaling operations and estimate the rescale factors using reciprocal singular value coefficients. Experimental results show the effectiveness of the algorithm under different rescale factors.     

An algorithm for the recovery of fetal heart rate series from CTG data

Cardiotocography (simultaneous recording of fetal heart rate (FHR) and uterine contractions) is one of the most used diagnostic techniques to evaluate fetal well-being and to investigate the functional state of the fetal autonomic nervous system. Recently, great interest has been paid to the variability of the FHR, and its frequency analysis, as a base for a more objective analysis of the cardiotocographic (CTG) tracings. FHR signals are unevenly sampled series. To obtain evenly sampled series, cardiotocographs often use zero-order interpolation. Such process is simple and fast but results unsuitable for frequency analyses because it introduces alterations in the FHR power spectrum. An algorithm for the recovery of the true FHR series out of the zero-order interpolated CTG data was developed and evaluated.     

A Higher Order Interpolation Scheme of Finite Volume Method for Compressible Flow on Curvilinear Grids

A higher order interpolation scheme based on a multi-stage BVD (Boundary Variation Diminishing) algorithm is developed for the FV (Finite Volume) method on non-uniform, curvilinear structured grids to simulate the compressible turbulent flows. The designed scheme utilizes two types of candidate interpolants including a higher order linear-weight polynomial as high as eleven and a THINC (Tangent of Hyperbola for INterface Capturing) function with the adaptive steepness. We investigate not only the accuracy but also the efficiency of the methodology through the cost efficiency analysis in comparison with well-designed mapped WENO (Weighted Essentially Non-Oscillatory) scheme. Numerical experimentation including benchmark broadband turbulence problem as well as real-life wall-bounded turbulent flows has been carried out to demonstrate the potential implementation of the present higher order interpolation scheme especially in the ILES (Implicit Large Eddy Simulation) of compressible turbulence.     

Release phenomena of insulin from an implantable device composed of a polyion complex of chitosan and sodium hyaluronate

An implant controlled-release system for protein drug delivery based on a polyion complex device composed of chitosan (CS) and sodium hyaluronate (HA) was investigated. The conditions which generated the greatest amount of the polyion solid complex were studied to ascertain the formation of polyion complex between CS and HA. The greatest amount of the polyion complex was formed at the weight ratio of 3 to 7 (CS:HA) at pH 3.5. Furthermore, the CS–HA pellets were prepared and then drug release from CS–HA pellets was evaluated using insulin as a model drug. The results demonstrated that the insulin release from CS–HA pellets was markedly influenced by both the change in the polymer mixing ratio and the total pellet weight, whereas the compression pressure did not affect the release significantly. An artificial neural network (ANN) and biharmonic spline interpolation (HSI) were employed to predict the actual relation between causal factors and the release rate constant of insulin. Although both the ANN and HSI successfully represented a non-linear relationship between the formulation factors and the release rate constant, HSI methodology gave a better estimation than that of the ANN.     

基于光电容积脉搏波的呼吸信号提取及其系统的研制

血氧饱和度和呼吸信号是人体重要的生理信号,呼吸的监测在临床以及生活中具有重要的意义.该研究建立了一套从光电容积脉搏波中提取呼吸信号的系统.包括脉搏波信号的采集、呼吸信号的提取以及从下位机中实时传输至上位机的呼吸率和脉率的计算.     

Assessment of plantar flexors activation capacity: nerve versus muscle stimulation by single versus double pulse

This study was designed to investigate if the relationship between the interpolated twitch-torque (IT) and voluntary torque (VT) is affected by the number of electrical stimuli (single vs. double) and the stimulation site (nerve trunk vs. muscle). The results showed that the IT–VT relationship of the plantar flexors is appropriately described by a composite (linear + curvilinear) model. Indeed, whatever the stimulation method, the IT–VT relationship was linear between approximately 25 and 75% of the maximal voluntary torque (MVT) and curvilinear for higher contraction intensities. The four stimulation conditions are equivalent in assessing the maximal voluntary activation (VA% range 96.2 ± 5.0 to 98.5 ± 3.1%) as well as in determining the true maximal torque expected for total twitch occlusion (MT_exp range 171.4 ± 21.2 to 179.0 ± 26.8 Nm). The gap between the MVT and MT_exp should be viewed as an index of muscle inactivation. This gap was comparable for the four stimulation methods (2–6%) and close to the deficit in VA% (2–4%). No pulse-number effect was found on the IT–VT relationship when the nerve was stimulated but an effect on the concavity of the composite relationship was observed when the stimulation was applied over the muscle. Even though the four stimulation techniques are equivalent in assessing the maximal activation capacity our results demonstrate that the neural stimulation method is the most consistent as it guarantees the same motor pool recruitment independently from the number of pulses.