基于心脏磁共振电影图像的压缩激励残差U形网络左心肌分割

左心肌分割对心脏疾病诊疗具有重要意义.但左心肌内部毗邻乳头肌、小梁,外部与周围组织灰度相近,是分割难点.本文首先对心脏磁共振电影图像数据进行感兴趣区域提取等预处理;其次,搭建融合了压缩激励模块和残差模块的U形网络(SERU-net)分割左心肌;最后,利用75例数据训练SERU-net网络,对18例数据进行预测.基于本文方法的分割结果相对于金标准的Dice系数与豪斯多夫距离均值分别是0.902、2.697 mm;利用本文方法分割得到的舒张末期、收缩末期左心室心肌质量与金标准的相关系数和偏差均值分别是0.995、0.993和3.784 g、2.338 g.结果表明,本文方法与金标准匹配程度较高,有望辅助诊断心脏疾病.     

外伸梁差分离散系统的模态反问题

研究外伸梁模态的反问题,即由两组位移模态或应变模态及相应的频率构造外伸梁差分离散模型物理参数的方法,讨论反问题解存在的充要条件.提出算法并用算例验证.结果表明,由应变模态构造外伸梁的物理参数比用位移模态构造梁的物理参数精度要高.     

基于COLLATE融合多图谱的心脏电影MRI右心室分割

右心室分割对肺动脉高压等疾病的心功能分析具有重要的临床意义.然而，右心室心肌薄、易变且不规则，其传统的医学图像分割方法仍然未能取得突破性进展.本文提出基于COLLATE（Consensus Level，Labeler Accuracy and Truth Estimation）的多图谱分割方法，首先以归一化互信息为相似测度对目标图像和图谱集进行B样条配准以获取粗分割结果；然后利用COLLATE对粗分割结果进行融合；最后采用基于形状约束的区域生长算法修正出现错误的数据.10例临床心脏磁共振短轴电影图像被用于算法验证.本文还将使用基于COLLATE的多图谱分割方法得到的结果与深度学习算法及手动分割进行了比较.结果显示与深度学习算法比较，使用本文算法得到的射血分数（Ejection Fraction，EF）与手动分割更加一致和相关，表明该算法的分割结果有望辅助临床心脏功能诊断.     

应变Si/(001)S1－xGex本征载流子浓度模型

利用应变Si CMOS技术提高载流子迁移率是当前研究发展的重点,本征载流子浓度是应变Si材料的重要物理参数,也是决定应变Si器件电学特性的重要参量.本文基于K.P理论框架,从分析应变Si/（001）Si_1-xGe_x材料能带结构出发,详细推导建立了300K时与Ge组分（x）相关的本征载流子浓度模型.该数据量化模型可为Si基应变器件物理的理解及器件的研究设计提供有价值的参考. 关键词： 应变Si 有效态密度 本征载流子浓度     

单轴应变硅nMOSFET栅隧穿电流模型

本文基于量子机制建立了单轴应变硅nMOSFET栅隧穿电流模型,分析了隧穿电流与器件结构参数、偏置电压及应力的关系.仿真分析结果与单轴应变硅nMOSFET的实验结果符合较好,表明该模型可行.同时与具有相同条件的双轴应变硅nMOSFET的实验结果相比,隧穿电流更小,从而表明单轴应变硅器件更具有优势.该模型物理机理明确,不仅适用于单轴应变硅nMOSFET,只要将相关的参数置换,该模型也同样适用于单轴应变硅pMOSFETs. 关键词： 单轴应变 nMOSFET 栅隧穿电流 模型     

数字图像相关中基于位移场局部最小二乘拟合的全场应变测量

为了从含噪声的位移场中计算得到可靠的应变场,提出一种基于位移场局部最小二乘拟合的全场应变求解方法。介绍了数字图像相关方法的原理,阐述了基于位移场局部最小二乘拟合的全场应变求解方法,并讨论了计算区域边界、孔洞及裂纹附近区域等情况下的应变计算。对均匀变形和中心带圆孔的薄铝板拉伸实验的计算结果表明,该方法能有效地从原始位移场数据中提取全场应变信息。在均匀变形情况下应选择大的应变计算窗口,计算结果更逼近真值;在非均匀变形情况下,如果位移场中包含较强的噪声,则应选择较大的应变计算窗口,而位移场精度很高时可选择更小的应变计算窗口。     

反向导引场自由电子激光器的三维非线性模拟

 对Conde–Bekefi反向导引场自由电子激光（FEL）放大器实验进行了三维非线性分析。当引入一类似于回旋自谐振脉塞收缩角(pinch angle)参数的电子束入射角参数后，其模拟计算结果与实验相符。这预示了反向导引场FEL中束波互作用的一种新机理。     

基于密集多尺度U-net网络的电影心脏磁共振图像右心室自动分割

右心室分割对心脏功能评估具有重要意义.然而,右心室结构复杂,传统分割方法效果欠佳.本文提出一种密集多尺度U-net（DMU-net）网络用于分割右心室,首先对56例数据进行归一化、增强及感兴趣区域提取的预处理;然后结合多尺度融合和嵌套密集连接结构搭建网络;最后利用预处理后的数据对DMU-net网络进行训练和验证,并对15例仅提取感兴趣区域的数据进行测试.本文方法与手动分割的Dice系数和豪斯多夫距离平均值分别为0.862和4.44 mm,优于文献中其它分割效果较好的方法;舒张末期容积、收缩末期容积、射血分数及每搏输出量的相关系数为0.992、0.960、0.987和0.982.结果表明,使用本文方法的分割结果与手动分割结果重合度高、差异性小,有望为心脏疾病诊断提供参考.     

用于多参数同时测量的散斑干涉仪

本文描述了一种新的散斑干涉仪,它可以同时对表面应变物体的面内位移、离面位移、位移梯度和莫尔曲率进行测量;由于利用了取向滤波方法,在滤波系统输出平面同时得到了与这些参数对应的散斑条纹.还利用散斑空间运动规律较好地解释了散斑干涉与散斑剪切干涉之间的关系.     

剪切应变下刃型位错的滑移机理的晶体相场模拟

针对刃型位错的滑移运动, 构建包含外力场与晶格原子密度耦合作用的体系自由能密度函数, 建立剪切应变作用体系的晶体相场模型. 模拟了双相双晶体系的位错攀移和滑移运动, 计算了位错滑移的Peierls势垒和滑移速度. 结果表明: 施加较大的剪切应变率作用, 体系能量变化为单调光滑曲线, 位错以恒定速度做连续运动, 具有刚性运动特征; 剪切应变率较小时, 体系能量变化出现周期波动特征, 位错运动是处于低速不连续运动状态, 运动出现周期“颠簸”式滑移运动, 具有黏滞运动特征; 位错启动运动, 存在临界的势垒. 位错启动攀移运动的Peierls势垒要比启动滑移Peierls势垒大几倍. 位错攀移和滑移运动特征与实验结果相符合.     

Cardiovascular magnetic resonance virtual tagging with B-spline-based free-form deformation

PurposeWe developed a virtual tagging technique that reconstructs tagging images using the displacement field obtained by applying B-spline free-form deformation (FFD) between diastolic images and images of other cardiac phases in cardiac cine MRI. The purpose of this study was to validate its characteristics and usefulness in phantom and patient studies.MethodsDigital phantoms simulating uniform and non-uniform wall motion models were created, and virtual tagging images were reconstructed with various matrix sizes and tag resolutions to evaluate the accuracy of FFD and the characteristics of the tags. In the patient study, FFD's accuracy was assessed at three levels (base, middle, and apex) in healthy patients. In patients with heart failure, virtual tagging images were compared with strain maps obtained by feature tracking and virtual tagging.ResultsIn the phantom study, blurring of tags was observed when tags were reconstructed with high resolution using a small matrix size. In the patient study, the accuracy of FFD was lower in the base than in the apex. Patients with heart failure had decreased distortion of the displacement field vector and virtual tags, indicating decreased local wall motion, consistent with areas of abnormalities found in strain maps.ConclusionThe virtual tagging technique does not require additional imaging and can visualize regional LV motion abnormalities via deformation of the tag as well as conventional cardiovascular magnetic resonance tagging.     

Hybrid segmentation of left ventricle in cardiac MRI using gaussian-mixture model and region restricted dynamic programming

Segmentation of the left ventricle from cardiac magnetic resonance images (MRI) is very important to quantitatively analyze global and regional cardiac function. The aim of this study is to develop a novel and robust algorithm which can improve the accuracy of automatic left ventricle segmentation on short-axis cardiac MRI. The database used in this study consists of three data sets obtained from the Sunnybrook Health Sciences Centre. Each data set contains 15 cases (4 ischemic heart failures, 4 non-ischemic heart failures, 4 left ventricle (LV) hypertrophies and 3 normal cases). Three key techniques are developed in this segmentation algorithm: (1) ray scanning approach is designed for segmentation of images with left ventricular outflow tract (LVOT), (2) a region restricted technique is employed for epicardial contour extraction, and (3) an edge map with non-maxima gradient suppression approach is put forward to improve the dynamic programming to derive the epicardial boundary. The validation experiments were performed on a pool of data sets of 45 cases. For both endo- and epi-cardial contours of our results, percentage of good contours is about 91%, the average perpendicular distance is about 2 mm. The overlapping dice metric is about 0.92. The regression and determination coefficient between the experts and our proposed method on the ejection fraction (EF) is 1.01 and 0.9375, respectively; they are 0.9 and 0.8245 for LV mass. The proposed segmentation method shows the better performance and is very promising in improving the accuracy of computer-aided diagnosis systems in cardiovascular diseases.     

Left ventricular 4D echocardiogram motion and shape analysis

The article describes the methodology and the processes of modeling the function and the motion of left ventricle using transesophageal echocardiograph. The parameters can be used in studying the functionality of left ventricle, the status of abnormality of myocardial, and the geometric and morphological of left ventricle in shape analysis. The parameters describes the motion of left ventricle are the left ventricular (LV) floating long axis, the morphological parameters. The LV morphological parameters describe the wall motion, the LV chamber cavity variation, the effective R-ratio of endomyocardial chamber of LV, the area surface curvature, and the global surface curvature circularity. The parameters such as stroke volume, ejection fraction used in evaluation of LV functions are also extracted.     

Dynamic pixel-wise weighting-based fully convolutional neural networks for left ventricle segmentation in short-axis MRI

Left ventricle (LV) segmentation in cardiac MRI is an essential procedure for quantitative diagnosis of various cardiovascular diseases. In this paper, we present a novel fully automatic left ventricle segmentation approach based on convolutional neural networks. The proposed network fully takes advantages of the hierarchical architecture and integrate the multi-scale feature together for segmenting the myocardial region of LV. Moreover, we put forward a dynamic pixel-wise weighting strategy, which can dynamically adjust the weight of each pixel according to the segmentation accuracy of upper layer and force the pixel classifier to take more attention on the misclassified ones. By this way, the LV segmentation performance of our method can be improved a lot especially for the apical and basal slices in cine MR images. The experiments on the CAP database demonstrate that our method achieves a substantial improvement compared with other well-know deep learning methods. Beside these, we discussed two major limitations in convolutional neural networks-based semantic segmentation methods for LV segmentation.     

Patterns of spiral tip motion in cardiac tissues

In support of the spiral wave theory of reentry, simulation studies and animal models have been utilized to show various patterns of spiral wave tip motion such as meandering and drifting. However, the demonstration of these or any other patterns in cardiac tissues have been limited. Whether such patterns of spiral tip motion are commonly observed in fibrillating cardiac tissues is unknown, and whether such patterns form the basis of ventricular tachycardia or fibrillation remain debatable. Using a computerized dynamic activation display, 108 episodes of atrial and ventricular tachycardia and fibrillation in isolated and intact canine cardiac tissues, as well as in vitro swine and myopathic human cardiac tissues, were analyzed for patterns of nonstationary, spiral wave tip motion. Among them, 46 episodes were from normal animal myocardium without pharmacological perturbations, 50 samples were from normal animal myocardium, either treated with drugs or had chemical ablation of the subendocardium, and 12 samples were from diseased human hearts. Among the total episodes, 11 of them had obvious nonstationary spiral tip motion with a life span of >2 cycles and with consecutive reentrant paths distinct from each other. Four patterns were observed: (1) meandering with an inward petal flower in 2; (2) meandering with outward petals in 5; (3) irregularly concentric in 3 (core moving about a common center); and (4) drift in 1 (linear core movement). The life span of a single nonstationary spiral wave lasted no more than 7 complete cycles with a mean of 4.6+/-4.3, and a median of 4.5 cycles in our samples. Conclusion: (1) Patently evident nonstationary spiral waves with long life spans were uncommon in our sample of mostly normal cardiac tissues, thus making a single meandering spiral wave an unlikely major mechanism of fibrillation in normal ventricular myocardium. (2) A tendency toward four patterns of nonstationary spiral tip motion was observed. (c) 1998 American Institute of Physics.     

Effect of hyperosmotic mannitol on magnetic resonance relaxation parameters in reperfused canine myocardial infarction

To determine how administration of a hyperosmotic agent alters regional nuclear magnetic resonance (NMR) relaxation parameters and imaging characteristics in ischemic-reperfused myocardium, 7 dogs were infused with mannitol for 15 minutes before and after the release of a 3 hour left anterior descending coronary artery (LAD) occlusion. Nine control animals received normal saline during the 3 hour occlusion and 1 hour reperfusion periods. Normal posterior left ventricular (LV) wall and the ischemic anterior LV wall (risk area) myocardium was sampled for calculation of segmental microsphere myocardial blood flow, % tissue water content, NMR relaxation times (T1, T2) and myocyte ultrastructure using electron microscopy. Mean infarct T1 values were 14% greater than normal segments in saline-treated controls, but only 5% greater after mannitol. The difference in tissue water content between infarcted and normal segments was 4% in saline-treated (83 vs. 79%) compared to 2% in mannitol-treated dogs (79 vs. 77%). T1, T2 and % water content of control infarct segments were greater than treated infarcts (p less than 0.01). T1 and T2 rose as occlusion flow fell below 0.5 ml/min/g in control hearts but did not rise until flows were reduced to 0.1 ml/min/g in mannitol-treated hearts. Areas of increased signal in T1 and T2 NMR images correlated well with histochemical infarct volume (r = 0.98, SEE = 1.1 cc) in mannitol-treated dogs, but infarct borders were qualitatively less well-defined than in controls. We concluded that mannitol (1) diminishes tissue edema and reduces NMR relaxation parameters (T1, T2) in infarcted myocardium; and (2) attenuates the rise in T1 and T2 and ultrastructural myocyte injury in ischemic-reperfused myocardium.     

Study of myocardial fiber pathway using magnetic resonance diffusion tensor imaging

The purpose of this study was to investigate myocardial fiber pathway distribution in order to provide supplemental information on myocardial fiber architecture and cardiac mechanics. Diffusion tensor imaging (DTI) with medium diffusion resolution (15 directions) was performed on normal canine heart samples (N=6) fixed in formalin. With the use of diffusion tensor fiber tracking, left ventricle (LV) myocardial fiber pathways and helix angles were computed pixel by pixel at short-axis slices from base to apex. Distribution of DTI-tracked fiber pathway length and number was analyzed quantitatively as a function of fiber helix angle in step of 9 degrees . The long fiber pathways were found to have small helix angles. They are mostly distributed in the middle myocardium and run circumferentially. Fiber pathways tracked at the middle and upper LV are generally longer than those near the apex. Majority of fiber pathways have small helix angles between -20 degrees and 20 degrees , dominating the fiber architecture in myocardium. Likely, such myocardial fiber pathway measurement by DTI may reflect the spatial connectiveness or connectivity of elastic myofiber bundles along their preferential pathway of electromechanical activation. The dominance of the long and circumferentially running fiber pathways found in the study may explain the circumferential predominance in left ventricular contraction.     

The mass normalization of the displacement and strain mode shapes in a strain experimental modal analysis using the mass-change strategy

The classic experimental modal analysis (EMA) is a well-known procedure for determining the modal parameters. The less frequently used strain EMA is based on a response measurement using strain sensors. The results of a strain EMA are the modal parameters, where in addition to the displacement mode shapes the strain mode shapes are also identified. The strain EMA can be used for an experimental investigation of a stress–strain distribution without the need to build a dynamical model. It can also be used to determine the modal parameters when, during modal testing, a motion sensor cannot be used and so a strain sensor is used instead. The displacement and strain mode shapes that are determined with the strain EMA are not mass normalized (scaled with respect to the orthogonality properties of the mass-normalized modal matrix), and therefore some dynamical properties of the system cannot be obtained. The mass normalization can be made with the classic EMA, which requires the use of a motion sensor. In this research a new approach to the mass normalization in the strain EMA, without using a motion sensor, is presented. It is based on the recently introduced mass-change structural modification method, which is used for the mass normalization in an operational modal analysis. This method was modified in such a way that it can be used for the mass normalization in the strain EMA. The mass-normalized displacement and strain mode shapes were obtained using a combination of the proposed approach and the strain EMA. The proposed approach was validated on real structures (beam and plate).     

Convexity preserving level set for left ventricle segmentation

In clinical applications of cardiac left ventricle (LV) segmentation, the segmented LV is desired to include the cavity, trabeculae, and papillary muscles, which form a convex shape. However, the intensities of trabeculae and papillary muscles are similar to myocardium. Consequently, segmentation algorithms may easily misclassify trabeculae and papillary muscles as myocardium. In this paper, we propose a level set method with a convexity preserving mechanism to ensure the convexity of the segmented LV. In the proposed level set method, the curvature of the level set contours is used to control their convexity, such that the level set contour is finally deformed as a convex shape. The experimental results and the comparison with other level set methods show the advantage of our method in terms of segmentation accuracy. Compared with the state-of-the-art methods using deep-learning, our method is able to achieve comparable segmentation accuracy without the need for training, while the deep-learning based method requires a large set of training data and high-quality manual segmentation. Therefore, our method can be conveniently used in situation where training data and their manual segmentation are not available.     

Left ventricle segmentation via two-layer level sets with circular shape constraint

This paper proposes a circular shape constraint and a novel two-layer level set method for the segmentation of the left ventricle (LV) from short-axis magnetic resonance images without training any shape models. Since the shape of LV throughout the apex-base axis is close to a ring shape, we propose a circle fitting term in the level set framework to detect the endocardium. The circle fitting term imposes a penalty on the evolving contour from its fitting circle, and thereby handles quite well with issues in LV segmentation, especially the presence of outflow track in basal slices and the intensity overlap between TPM and the myocardium. To extract the whole myocardium, the circle fitting term is incorporated into two-layer level set method. The endocardium and epicardium are respectively represented by two specified level contours of the level set function, which are evolved by an edge-based and a region-based active contour model. The proposed method has been quantitatively validated on the public data set from MICCAI 2009 challenge on the LV segmentation. Experimental results and comparisons with state-of-the-art demonstrate the accuracy and robustness of our method.