基于特征约束点的纹理映射算法

纹理映射技术用于生成物体表面的纹理细节,是真实感图形技术的重要组成部分,也是计算机图形学的一个重要研究方向.针对目前很多纹理映射算法计算量大,方法比较复杂的缺点,应用Candide3作为三维网格模型,提出了一种快速有效的基于特征约束点的纹理映射算法.通过在三维网格模型和纹理图像上选取少量对应的特征约束点,利用三角网格剖分算法在纹理图像上建立选取特征点的三角网格.进而通过求取质心坐标的方法计算出三维网格模型上所有特征点的纹理坐标并完成整个三维网格模型的纹理映射.实验结果表明,提出的算法计算速度较快,能够得到高真实度的纹理映射效果,并且适用于不同纹理图像映射到同一三维网格模型上.     

基于调和映射的约束纹理映射方法

传统的约束纹理映射方法大都建立在迭代优化的基础上,给出的解多为近似解．为此,提出了一种基于调和映射的约束纹理映射方法,利用该方法可以得到约束纹理映射问题的一个形式化精确解．由于调和映射具有保持映射能量最小的良好性质,因此该方法能够最小化纹理映射的形变;另外,约束的纹理映射是个大交互量的工作,埘映射效果的优化调整非常重要,提出的自适应局部邻域调整方法能够实现映射效果的实时优化．该方法鲁棒并且效率高,实验结果表明利用该方法能够取得良好的绘制效果．     

3D geometry-dependent texture map compression with a hybrid ROI coding

Difering from common 2D images,a texture map,since it is used to project onto a 3D model in 3D space,not only contains 2D texture information,but also implicitly associates certain 3D geometric information.Related to this,an efective 3D geometry-dependent texture map compression method with hybrid region of interest(ROI)coding is proposed in this paper.We regard the visually important area of the texture map as the ROI.To acquire the visually important areas of the texture map,we take into account information from both the 3D geometry and 2D texture maps,depicting the saliency of the textured model,distortion of the texture mapping,and boundary of the texture atlas.These visually important areas are expressed as a visual importance map.According to the particularity of the texture map,a hybrid ROI coding method that utilizes Max-Shift and an improved post compression rate distortion(PCRD)technique is presented,guided by this visual importance map.To find the exact wavelet coefcients pertaining to these ROIs before carrying out the hybrid ROI coding,this paper proposes a stochastic coefcient priority mask map computational method.Experimental results show that the visually important areas of the texture image have a better visual efect and that a good rendering result can be obtained from the texture mapping.     

Improved two-way satellite time and frequency transfer with Multi-GEO in BeiDou navigation system

Due to the limitation of Geostationary Earth Orbit(GEO)satellite resources,only very few TWSTFT links with parallel counterpart operating via diferent GEO satellites between two ground stations are available.There are 5 GEO satellites in BeiDou Navigation System,which enables an improved TWSTFT with Multi-GEO.Two algorithms are put forward to take the advantage of Multi-GEO links to improve the performance of time and frequency transfer respectively.The dynamic weight average algorithm based on a newly proposed dynamic time variance aiming at improving the performance of time transfer is shown to be able to reduce the dynamic time deviation of the two-way measurement result to 1/√N dynamic time deviation of the single link using N-GEO links,if each link has similar level of link noises.The algorithm of cancelling out the link noise Allan variances with 2-GEO links is expected to obtain the Allan variance of clock diference between two ground stations.The experiment results using 2-GEO links in BeiDou Navigation System show that the dynamic weight average algorithm can improve the dynamic time deviation of the two-way measurement to around 0.6 of that of the single link in an average concept.And the estimated Allan deviation of the clock diference using the algorithm for cancelling the noise Allan variances of both links can reach 7×10 12(τ0.65)for the averaging time 1–105s which is within the same order of magnitude as the estimated Allan deviation based on carrier Doppler measurement for the averaging time 10 s–104s,despite the noisy links with the Allan deviation follows 3×10 10(τ1).     

A PC-based high-quality and interactive virtual endoscopy navigating system using 3D texture based volume rendering

As an alternative method to optical endoscopy, visual quality and interactivity are crucial for virtual endoscopy. One solution is to use the 3D texture map based volume rendering method that offers high rendering speed without reducing visual quality. However, it is difficult to apply the method to virtual endoscopy. First, 3D texture mapping requires a high-end graphic workstation. Second, texture memory limits reduce the frame-rate. Third, lack of shading reduces visual quality significantly. As 3D texture mapping has become available on personal computers recently, we developed an interactive navigation system using 3D texture mapping on a personal computer. We divided the volume data into small cubes and tested whether the cubes had meaningful data. Only the cubes that passed the test were loaded into the texture memory and rendered. With the amount of data to be rendered minimized, rendering speed increased remarkably. We also improved visual quality by implementing full Phong shading based on the iso-surface shading method without sacrificing interactivity. With the developed navigation system, 256 x 256 x 256 sized brain MRA data was interactively explored with good image quality.     

基于三维几何视觉重要性的纹理图像选择压缩算法

不同于传统二维图像,映射到三维模型上的纹理图像隐式包含了三维几何视觉信息. 然而,目前已有的纹理图像压缩方法并未考虑此特性. 本文提出了一种与三维模型几何视觉特性相关的纹理图像选择压缩算法. 首先给出一种结合纹理图像的显著性及其纹理走样的视觉重要性图构建方法, 将纹理图像划分为具有不同优先级别区域.之后,利用提出的选择压缩方法对它们进行不同比例压缩. 实验结果表明当选择本压缩算法时,纹理化三维模型能够获取较好的视觉效果.     

三维模型表面特征的水墨效果渲染算法

提出一种对三维模型实现水墨效果渲染的两步二维纹理映射方法.该方法首先通过模型顶点的视线向量和法线向量的点乘积计算其夹角的余弦值,作为第一步的纹理坐标进行映射;再通过欧拉公式计算径向曲率及其方向导数值,作为第二步映射的纹理坐标.为达到实时的交互浏览,在第二步映射时还可采用基于视点无关的曲率特征的映射策略(如平均曲率、高斯曲率、主曲率等).在快速生成水墨画效果的同时也显示出更多的特征细节.     

基于几何与图像混合绘制中的快速WARP变换算法研究

基于几何与图像的混合绘制中，3D Warp算法以严格的数学变换为基础，从而能够保证准确的投影关系，但该算法在实时绘制阶段需进行大量的数学运算，故其时间复杂度较高，该文提出了一种新的快速Warp变换算法，算法以3D Warp算法为基础，采用了崭新的投影过程，从而使时间复杂度较3D Warp算法有较大幅度的下降（降低约7．51倍）。同时，该算法是一种流水结构，能够有效利用现有的加速硬件，而无需改变图形硬件的体系结构。     

Layered textures for image-based rendering

An extension to texture mapping is given in this paper for improving the efficiency of image-based rendering. For a depth image with an orthogonal displacement at each pixel, it is decomposed by the displacement into a series of layered textures (LTs) with each one having the same displacement for all its texels. Meanwhile, some texels of the layered textures are interpolated for obtaining a continuous 3D approximation of the model represented in the depth image. Thus, the plane-to-plane texture mapping can be used to map these layered textures to produce novel views and the advantages can be obtained as follows: accelerating the rendering speed, supporting the 3D surface details and view motion parallax, and avoiding the expensive task of hole-filling in the rendering stage. Experimental results show the new method can produce high-quality images and run faster than many famous image-based rendering techniques.     

纹理映射中的平面校正技术研究

为了快速实时地进行由平面组成的结构景物的3D建模问题,文中介绍了一种在进行图像3D重构时纹理映射中的平面校正方法。介绍了三角形模型在图像处理、图形绘制、虚拟现实等技术中的重要作用。从射影几何的角度出发,给出了从两幅视图进行景物三维重构的分层重构方法。在已知欧氏重构即摄像机内参数的基础上,介绍一种基于标定的平面射影失真矫正方法。通过此方法,将矫正过的纹理映射到欧式点重构结构中,得到景物的3D模型。经实验验证,这种方法在处理由平面组成的景物的3D重构中是实时有效的。     

Compression-Based 3D Texture Mapping for Real-Time Rendering

While 2D texture mapping is one of the most effective of the rendering techniques that make 3D objects appear visually interesting, it often suffers from visual artifacts produced when 2D image patterns are wrapped onto the surfaces of objects with arbitrary shapes. On the other hand, 3D texture mapping generates highly natural visual effects in which objects appear carved from lumps of materials rather than laminated with thin sheets as in 2D texture mapping. Storing 3D texture images in a table for fast mapping computations, instead of evaluating procedures on the fly, however, has been considered impractical due to the extremely high memory requirement. In this paper, we present a new effective method for 3D texture mapping designed for real-time rendering of polygonal models. Our scheme attempts to resolve the potential texture memory problem by compressing 3D textures using a wavelet-based encoding method. The experimental results on various nontrivial 3D textures and polygonal models show that high compression rates are achieved with few visual artifacts in the rendered images and a small impact on rendering time. The simplicity of our compression-based scheme will make it easy to implement practical 3D texture mapping in software/hardware rendering systems including real-time 3D graphics APIs such as OpenGL and Direct3D.     

一种改进的人脸纹理映射方法

识别人脸图像优化问题,由于人脸纹理映射时容易出现的纹理扭曲和畸变现象,影响图像的真实性,提出一种改进的纹理图像映射方法。根据人脸特征点信息,利用特征块匹配法配准正侧面人脸图像,并使用加权平滑算法和金字塔方法对图像进行双重融合实现图像的平滑过渡;然后利用一种特殊的定位方法来计算三维人脸模型上顶点对应的纹理坐标;最后根据纹理坐标将纹理图像映射到三维特定人脸模型表面上,可得到具有真实感的三维人脸模型。实验结果表明,纹理合成方法简捷有效,纹理映射结果生动逼真,有效的避免了纹理图像的扭曲和畸变现象,具有较高的鲁棒性。     

基于面部特征约束的人脸纹理映射及变形

将一幅人脸图像（称为源图像）映射到另一幅人脸图像（称为目标图像）上,从而实现人脸图像的变形并达到某种特殊效果的表现要求。其中,一个最关键的问题是保持两幅人脸中器官特征的约束映射。提出一种技术策略,可以有效实现这种映射效果。利用人脸特征检测算法检测出源人脸和目标人脸的特征点;基于特征点,利用Delaunay三角剖分的方法将目标人脸转化为三角网格;采用调和映射的方法计算得出所有网格点在源人脸图像中的对应坐标;在获得网格点纹理坐标后,就可以实现源人脸到目标人脸的映射。实验表明,该方法可以实现任意人脸间的映射,并能较好地保持人脸五官间的映射关系。     

虚平面映射:深度图像内部视点的绘制技术

首先推导与归纳了图像三维变换中像素深度场的变换规律,同时提出了基于深度场和极线原则的像素可见性别方法,根据上述理论和方法,提出一种基于深度图像的建模与绘制(image-based modeling and rendering,简称IBMR)技术,称为虚平面映射.该技术可以基于图像空间内任意视点对场景进行绘制.绘制时,先在场景中根据视线建立若干虚拟平面,将源深度图像中的像素转换到虚平面上,然后通过对虚平面上像素的中间变换,将虚平面转换成平面纹理,再利用虚平面的相互拼接,将视点的成像以平面纹理映射的方式完成.新方法还能在深度图像内侧,基于当前视点快速获得该视点的全景图,从而实现视点的实时漫游.新方法视点运动空间大、存储需求小,且可以发挥图形硬件的纹理映射功能,并能表现物体表面的三维凹凸细节和成像视差效果,克服了此前类似算法的局限和不足.     

复杂背景下彩色数字图像分割算法研究

为了提高复杂背景下彩色数字图像分割的准确率,减少噪声对图像处理的干扰,尽可能的缩短运行时间,需要对复杂背景下彩色数字图像分割算法进行研究。当前图像分割算法在彩色数字图像分割的过程中,仅仅考虑了图像像素的亮度值,没有考虑其空间特征,存在计算的复杂性过大等缺陷,影响彩色数字图像处理效果。为此,提出了一种基于随机权重粒子群的复杂背景下彩色数字图像分割算法。该算法先采用多尺度均匀滤波方法,对复杂背景下彩色数据图像进行划分,其中包含噪声和不含噪声的像素点的亮度值、结构元素以及局部区域内的图像像素加权亮度密度特征。采用多段图分割获取彩色数字图像的优化分割,在平滑项中代入彩色数字图像梯度信息,对彩色数字图像分割结果中的弱边界进行剔除,从而实现准确的彩色数字图像分割。实验仿真证明,所提算法增加了彩色数字图像分割的对比度和信噪比,提高了复杂背景下彩色数字图像分割的准确性。     

匹配图像与3维模型特征点的真实感3维头重建

提出并实现一种基于两张正交图像和一个标准3维头模型,并利用2D图像特征点和3D模型特征点的匹配进行3维头模型重建的算法。首先,进行面部区域和头发区域的分割,利用色彩传递对输入图像进行颜色处理。对正面图像利用改进后的ASM（主动形状模型）模型进行特征点定位。改进局部最大曲率跟踪(LMCT)方法,更为鲁棒的定位了侧面特征点。在匹配图像特征点与标准3维头上预先定义的特征点的基础上,利用径向基函数进行标准头形变,获得特定人的3维头部形状模型。采用重建好的3维头作为桥梁,自动匹配输入图像,进行无缝纹理融合。最后,将所得纹理映射到形状模型上,获得对应输入图像的特定真实感3维头模型。     

单幅人脸图像的全景纹理图生成方法

目的 针对从单幅人脸图像中恢复面部纹理图时获得的信息不完整、纹理细节不够真实等问题,提出一种基于生成对抗网络的人脸全景纹理图生成方法。方法 将2维人脸图像与3维人脸模型之间的特征关系转换为编码器中的条件参数,从图像数据与人脸条件参数的多元高斯分布中得到隐层数据的概率分布,用于在生成器中学习人物的头面部纹理特征。在新创建的人脸纹理图数据集上训练一个全景纹理图生成模型,利用不同属性的鉴别器对输出结果进行评估反馈,提升生成纹理图的完整性和真实性。结果 实验与当前最新方法进行了比较,在CelebA-HQ和LFW （labled faces in the wild）数据集中随机选取单幅正面人脸测试图像,经生成结果的可视化对比及3维映射显示效果对比,纹理图的完整度和显示效果均优于其他方法。通过全局和面部区域的像素量化指标进行数据比较,相比于UVGAN,全局峰值信噪比（peak signal to noise ratio,PSNR）和全局结构相似性（structural similarity index,SSIM）分别提高了7.9 dB和0.088,局部PSNR和局部SSIM分别提高了2.8 dB和0.053;相比于OSTeC,全局PSNR和全局SSIM分别提高了5.45 dB和0.043,局部PSNR和局部SSIM分别提高了0.4 dB和0.044;相比于MVF-Net （multi-view 3D face network）,局部PSNR和局部SSIM分别提高了0.6和0.119。实验结果证明,提出的人脸全景纹理图生成方法解决了从单幅人脸图像中重建面部纹理不完整的问题,改善了生成纹理图的显示细节。结论 本文提出的人脸全景纹理图生成方法,利用人脸参数和网络模型的特性,使生成的人脸纹理图更完整,尤其是对原图不可见区域,像素恢复自然连贯,纹理细节更真实。     

基于标准肤色的人脸图像纹理合成与三维重建应用

本文提出一种基于单幅人脸图像并结合标准肤色的人脸图像纹理合成和三维重建算法.首先,利用ASM算法提取人脸特征点,并通过基于局部线性嵌入算法的编辑传播实现颜色转换,使图像人脸色调与三维人脸模型标准肤色一致.接着,将人脸图像五官区域与标准肤色图进行泊松融合,并考虑眉毛遮挡情况,利用人脸对称性或眉毛模板还原眉毛.尤其对于半遮挡眉毛,采用Li模型和角点检测相结合的方法重建眉毛轮廓,得到最终人脸纹理图.最后通过纹理映射将人脸纹理图映射到三维人脸模型上,得到较好的个性化三维人脸重建效果.实验表明,本文算法能够适用于不同复杂背景和光照条件下拍摄的人脸图像,具有较快的处理速度,能够应用于人脸实时重建产品中.     

高质量的三维纹理硬件体绘制

与光线投射法相比,传统的3D纹理体绘制算法通常难以产生高质量的图像。为了增强渲染图像的真实感与质量,在基于GPU（Graphics Processing Unit）的三维纹理体绘制过程中以交互的速率实现了体阴影效果,并考虑现实图像合成中的可视化感知,提出将基于GPU的高动态范围色调映射技术应用到体绘制得到的结果图片中。最后对一些体数据集进行绘制,实验表明这些技术较好地解决了传统纹理绘制方法的缺点,提高了图像的质量。