基于图像分割的三维重构

三维重构方法是医学图像可视化系统、治疗计划系统的重要技术。基于图像分割的三维重构方法结合了图像分割、等值面抽取、网格简化三种技术,是不同于传统Marching Cubes算法的一种三维重构方法。它首先将医学图像分割为二值图,然后利用Marching Cubes方法进行等值面抽取,最后对得到的网格模型进行简化。实验结果表明,基于图像分割的三维重构方法加快了Marching Cubes的运算速度,改善了重构的效果,有利于实现对基于三维重构的大型几何模型的实时绘制和交互。     

基于近平面合并的三角网格简化算法

提出一种基于近平面合并的三角网格简化算法,针对三维重建MarchingCubes算法中产生的大量三角片进行简化。首先依据近平面判断条件将近似共面的三角形合并成一个大的多边形,然后按相应的准则将其重新进行三角化得到更少的三角形面片,以达到简化的目的。该算法在最大化的保留图像细节的前提下,精简了三角片的数目,提高了重建后图像绘制和传输的速率,节省了存储空间和处理时间。     

针对脊柱CT图像三维重构MC方法的改进

对原始MC（Marching Cubes）方法存在的问题和局限进行改进,并以此为基础重构脊柱三维模型,使之满足术前分析和远程手术的要求。采用双曲渐近线判别式消除面二义性问题,利用插值函数的空间连续性解决体二义性问题。提出基于对象相关性的cube连通性判断,减少三角面片过多的问题,进一步采用Decimation算法消减面片网格。实验证明了上述改进方法的有效性和正确性,可得到良好的三维重构效果。     

Synthesis and characterization of shaped ZnS nanocrystals in water in oil microemulsions

Shaped zinc sulfide nanocrystals were synthesized in W/O microemulsions by using cyclohexane/Triton X-100/n-pentanol/water system. Under different synthetic conditions appearance of two distinct morphologies of ZnS nanocrystals, either cubes or nanowires, was proven by transmission electron microscopy (TEM). The ZnS cubes have an average size of about 25 nm, while the ZnS nanowires have 25 Å diameter and length ranging from several hundred nanometers up to a few microns. The X-ray diffraction analysis (XRD) revealed formation of ZnS with cubic zinc blende crystal structure. Due to two dimensional confinement the exciton of ZnS nanowires is blue shifted compared to the bulk material. Four well-resolved photoluminescence bands in visible spectral region were observed upon excitation of cubic ZnS particles, while in the case of ZnS nanowires emission band was observed at 421 nm. The origin of photoluminescence bands was discussed in details.     

Electron beam induced synthesis of CdSe nanomaterials: Tuning of shapes from rods to cubes

Cadmium selenide (CdSe) nanomaterials of different shapes have been synthesized in the water pool of the water-in-oil microemulsions through 7 MeV electron beam irradiation. The rod shaped CdSe nanomaterials of lengths and breadths up to 1 μm and 200 nm, respectively with aspect ratios from 3 to 5 were grown in the case of microemulsions containing lower water content, water to surfactant concentrations ratio, w₀ = 10. In contrast, the cubic shaped CdSe nanomaterials of dimensions about 100 nm were formed in the microemulsions with higher water contents, w₀ values from 20 to 40. Such a transformation of shape from rod to cube was attributed to the change in the hydrophobicity and hydrophilicity of the micro-environment in the microemulsions. The as-grown CdSe nanomaterials were stable at ambient conditions and thus expected to be very useful in the device applications.     

实时模拟固体燃烧现象

为了模拟固体燃烧时产生的火焰及其伴随的物质分解过程,提出一个实时的燃烧模型.为了在模拟速度和视觉效果间取得一个很好的折中,采用一套混合的网格模拟系统;整个空间的热传导过程在一个统一的框架下模拟,因此能够很好地耦合动态的火焰和燃烧的固体.提出了一个改进的燃面更新方案以及一个新的固体表面纹理映射算法,能够生成逼真的结果.还提出了一系列提高模型效率的加速算法.最终实时地模拟出了真实感的固体燃烧现象.     

应用于一种工业单层CT的三维成像的实现方法

应用于集装箱、箱包CT检测系统的三维成像方法是此类工业CT检测系统产品化的重要研究部分。参考其它领域的CT三维成像方法,提出了三种可行性较大的成像算法,轮廓线连接法Marching Cubes(MC)方法以及基于距离函数的MC方法,并加以比较。对于集装箱、箱包CT检测系统的三维成像功能研制具有重要意义。     

Exact Isosurfaces for Marching Cubes

In this paper we study the exact contours of a piecewise trilinear scalar field. We show how to represent these contours exactly as trimmed surfaces of triangular rational cubic Bézier patches. As part of this, we introduce an extension of the marching cubes algorithm which gives a topologically exact triangular approximation of the contours for any case. Finally, we modify the exact contours to be globally  G¹  continuous without changing their topologies. We test the algorithm on both theoretical and practical data sets.     

基于Marching Cubes重组的外存模型渐进压缩

外存模型是指其规模远远超出内存容量的海量模型．为提高其存储、传输、显示等操作的效率，对外存模型进行渐进式的压缩是非常重要的．但当前已有的外存模型压缩算法都是单一层次的，不能做到渐进压缩．为此，该文提出一种针对外存模型的渐进压缩方法．能高效地压缩外存模型，并进行多分辨率的传输和显示．该方法首先将外存模型的包围盒空间按照八叉树形式进行剖分和层次化组织，使得最精细层次的各个立方块空间中的局部模型都能完全装入内存进行处理；然后，在各个立方块中对局部的模型进行基于Marching Cubes方式的重新拟合，并在此基础上建立各个局部的自适应八叉树；最后，基于各个局部自适应的八叉树，由粗至细渐进地遍历全局自适应八叉树的各个节点，并利用对内存模型能高效渐进压缩编码的先进方法进行编码压缩．实验表明，该方法对外存模型的压缩比达到了与处理内存模型相似的压缩比．高于目前的外存模型压缩方法，是第一个能渐进压缩外存模型的方法．     

Lossless binary image compression using logic functions and spectra

A lossless compression of images using coding schemes and patterns that include minterm, cube and coordinate data coding, Walsh, triangular and Reed-Muller weights based patterns, Reed-Muller spectra and reference row technique is proposed. The experimental results indicate that the technique is fairly efficient when compared with other methods based on representations of logic functions.