一种360度柱面全景图快速生成算法

全景图像是构造虚拟环境的基本要素之一,本文提出了一种新的360柱面全景图绘制算法。首先对原始图像进行柱面投影,然后通过对图象重叠部分的对应特征区域进行匹配和拼合来实现图象无缝镶嵌。实验表明该算法运算量较小,能有效、快速地生成柱面全景图。     

图像漫游：双灭点的松弛蜘蛛网格算法

为了解决蜘蛛网格算法在图像漫游中局限性，提出了双灭点的松弛蜘蛛网格算法。该算法在用于图像实时漫游时，用双灭点进行计算，其与前人提出的方法相比，更具通用性，不仅可以解决单灭点、双灭点、多灭点和灭线等不同类型的照片漫游问题，而且减轻了用户交互的工作量。此外，为了进一步提高图像的绘制质量，以获得高质量的漫游效果，还采用了一种分片绘制的策略。多种类型的图片实验证明了该算法的有效性。     

基于图像的马赛克

提出并实现了一种有效利用视觉马赛克现象来实现用较小图像自动拼合较大图像的图像合成方法，讨论了用图像合成马赛克图像的原理与模型，针对不同的实现目标给出了不同的实现算法，在总结实验结果的基础上展望了进一步研究的问题与方向。     

高质量图像绘制的LDI改进算法

对于漫反射场景,LayeredDepthImage(LDI)能够产生很好的绘制效果,但对于高光的场景,它存在绘制图像失真的问题。文章提出一种能够有效地绘制漫反射和高光场景的LDI改进算法。该算法对LDI的每个深度像素增加一个方向向量,对位于同一深度的像素采用加权平均的方法生成新视点下的目标图像。同时采用将方向向量离散化的方法,有效地降低了存储量。实验表明,新算法在绘制速度和绘制质量方面都取得了比较理想的效果。     

基于改进的Harris和SIFT算法的眼底图像拼合

传统的眼底图像拼合主要采用SIFT算法实现,算法计算量大,效率低,准确率也不高.Harris提取的是图像的角点信息,能较好准确地反应图像的特征信息.作者采用Harris和SIFT算法相结合的方式,采用Harris算法提取的角点信息,并用这些角点提取SIFT特征描述符,且在图像拼合的过程中对算法做出相应的改进,克服了多模式图像中局部梯度信息不一致带来的误差,计算效率高,且拼合效果较好.     

基于Delaunay三角网的模型拼合算法研究

模型拼合在地理建模、铁(公)路路线三维设计等领域应用十分广泛，研究如何快速高效地获得拼合模型的算法很有必要。基于Delaunay三角网模型的拼合算法，提出了一种快速获取模型拼合交线的方法和快速搜索交线范围内三角形的算法，采取拼合交线入网及初始D-三角网的裁剪2个关键步骤来实现D-三角网模型拼合。对模型拼合的整个过程进行了详细的阐述，采用VC^ 6．O语言实现了算法，并利用实验数据对算法进行测试，验证了算法的正确性与可行性。     

一种基于F求精的摄像机自定标方法

基于图像的建模与绘制(IBMR)是一种新型的真实感图形绘制方法，其中的关键技术如图像的立体匹配(Stereo matching)、三维重构(3D Reconstruction)和Warping绘制等，都以摄像机定标为前提．本文讨论一种基于F求精的自定标方法，在论述其基本原理的基础上，给出了相关算法与实现步骤，并用模拟照片的自定标计算结果和实拍照片的Warping绘制结果对算法进行了检验。     

一种光照鲁棒的图像拼接融合算法

针对传统图像拼接方法只能处理光照一致图像的问题,提出了一种对环境光照鲁棒的全景图拼接算法。该算法首先使用圆环投影来获取待拼接图像的匹配特征序列,不仅克服了传统图像特征提取方法中的区域局限性问题,而且较好地实现了光照变化的图像匹配;然后使用统计参数来调整待拼接图像的整体亮度,以解决光照变化问题;最后对于传统图像融合处理中采用线性加权函数通常引起的最终拼合图像重叠区域模糊问题,构造了包含图像梯度的能量函数,用于计算重叠区域的全局最优融合因子。实验表明,该算法对光照变化图像的拼接融合能取得满意的视觉效果。     

基于纹理合成的水墨山水画自动绘制

通过分析中国山水画的一般创作技法,提出了基于纹理合成的水墨山水画绘制方法.将一幅图像作为控制图像,用于控制要绘制的区域轮廓和明暗色调;采集画家的手稿作为样本图像,通过纹理合成绘制出国画绘画中的各种皴、染笔触;在传统的纹理合成方法上,提出了一种目标控制的多级纹理合成算法,该算法能利用控制图像有效地控制纹理合成的结果.从而可以将一幅数字照片或简单的手绘图自动地转换成各种中国水墨画效果的图像.仿真结果证明了该方法对于各种山水画的绘制具有良好的效果.     

基于SIFT特征匹配的图像拼接算法

基于SIFT特征匹配思想 ,提出一种快速、准确的图像拼接算法.SIFT特征匹配算法是目前国内外特征点匹配研究领域的热点与难点,其匹配能力较强,可以处理两幅图像之间发生平移、旋转、仿射变换情况下的匹配问题,甚至在某种程度上对任意角度拍摄的图像也具备较为稳定的特征匹配能力.在这一匹配思想的基础上结合多分辨率样条提出一种新的图像拼接算法,它采用多分辨率样条来确定拼接区域宽度和加权函数,使图像在不同分辨率下分别拼合,这样使拼合出的图像既清晰又光滑无缝.算法在大多数情况下能自动完成,实验结果证明该方法提高了拼接速度,降低了图像亮度差异的影响,使结果图像自然逼真.     

Rendering techniques for mixed reality

In mixed reality (MR) design review, the aesthetics of a virtual prototype is assessed by integrating a virtual model into a real-world environment and inspecting the interaction between the model and the environment (lighting, shadows and reflections) from different points of view. The visualization of the virtual model has to be as realistically as possible to provide a solid basis for this assessment and interactive rendering speed is mandatory to allow the designer to examine the scene from arbitrary positions. In this article we present a real-time rendering engine specifically tailored to the needs of MR visualization. The renderer utilizes pre-computed radiance transfer to calculate dynamic soft-shadows, high dynamic range images and image-based lighting to capture incident real-world lighting, approximate bidirectional texture functions to render materials with self-shadowing, and frame post-processing filters (bloom filter and an adaptive tone mapping operator). The proposed combination of rendering techniques provides a trade-off between rendering quality and required computing resources which enables high quality rendering in mobile MR scenarios. The resulting image fidelity is superior to radiosity-based techniques because glossy materials and dynamic environment lighting with soft-shadows are supported. Ray tracing-based techniques provide higher quality images than the proposed system, but they require a cluster of computers to achieve interactive frame rates which prevents these techniques from being used in mobile MR (especially outdoor) scenarios. The renderer was developed in the European research project IMPROVE (FP6-IST-2-004785) and is currently extended in the MAXIMUS project (FP7-ICT-1-217039) where hybrid rendering techniques which fuse PRT and ray tracing are developed.     

Illuminating the fourth dimension

A family of techniques for creating intuitively informative shaded images of 4-D mathematical objects is proposed. The rendering of an object in a 4-D world is described by considering step-by-step how objects might be rendered into images in simpler worlds. The mathematical principles needed to compute projected images of objects and their shadows in D dimensions are outlined. The issues involved in producing shaded images of objects in four dimensions, including extending rendering from 3-D to 4-D, smooth shading, and specularity, are discussed. Results of rendering a Steiner surface, torus, and knotted sphere in four dimensions are presented     

基于图像的光照模型研究综述

从传统图形学的绘制技术与基于图像的绘制技术相结合的角度出发,以全光函数这个基于图像的绘制技术的理论基础为核心,概括性地提出基于图像的光照研究的基本任务实际上是对全光函数的采样、重建、合成和重采样的过程,并进一步地指出,基于图像的光照研究的重要意义在于扩展了原有基于图像的绘制技术中只能改变视点位置和视线方向的限制,使之可以通过改变场景本身的组成成分产生出更加丰富的光照效果。同时,该文综述性地分析了近期内有关基于图像的光照问题的部分研究工作,并从如何改变场景光照条件的角度出发,按照所使用的光照模型的不同,将这些方法分成三大类,即利用传统光照模型的方法、利用基于图像的光照模型的方法以及无需光照模型的方法。并从这个分类框架出发,进一步分析指出,利用基于图像的光照模型的方法将是未来研究的重点,并沿着这一方向尝试性地提出了一种新的模型。     

基于全景图像的拼接算法研究

IBR技术不仅可以弥补传统基于几何绘制技术的不足,而且能给出更丰富的图像显示.论文从工程实际应用的角度出发,提出并实现了一种图像拼接的新方法.该方法即可用于全景控件图像的拼接,也可用于普通图像的拼接.试验证明,此方法在全景图像的生成中具有自动拼接、效率高等优点.     

How to Render Frames and Influence People

Rendering systems generally treat the production of images as an objective process governed by the laws of physics. However, perception and understanding on the part of viewers are subjective processes influenced by a variety offactors. For example, in the presentation of architectural drawings, the apparent precision with which the drawings are made will affect whether the viewer considers the design as part of a preliminary design or as part of a final polished project, and to some extent the level of confidence the viewer has in the encoded information. In this paper we develop techniques for rendering images in a way that differs from the usual photorealistic or wire-frame output of renderers. In particular, our techniques allow a user to adjust the rendering of a scene to produce images using primitives with variable degrees of precision, from approximations that resemble vague “five-minute-sketches” to more mature but still hand-drawn images. We provide a theoretical framework for analysing the information flow from the computer to the user via such images. Finally, we describe the design and implementation of a prototypical renderer and show examples of its output.     

基于二值体金字塔数据结构的自适应快速体绘制方法

直接体绘制技术无需拟合中间几何图元直接绘制体数据,由于绘制图像质量高,该技术倍受青睐。但是,因为直接体绘制需要遍历整个数据场,绘制速度慢,在绘制大规模数据场时,速度已经成为应用该项技术的一大瓶颈。为此,提出一种基于二值体金字塔数据结构的自适应快速体绘制方法,该方法可以在不降低图像质量的情况下,加快图像绘制速度。     

Artistic rendering enhancing global structure

Non-photorealistic rendering techniques usually produce abstracted images. Most existing methods consider local rendering primitives, and global structures may be easily obscured. Inspired by artists, we propose a novel image abstraction method that considers preserving or even enhancing global structures in the input images. Linear structures are particularly considered due to their wide existence and the availability of techniques for their reliable detection. Based on various computer vision techniques, the algorithm is fully automatic. As demonstrated in the paper, artistic looking results are obtained for various types of images. The technique is orthogonal to many non-photorealistic rendering techniques and can be combined with them.     

Two-level volume rendering

Presents a two-level approach for volume rendering, which allows for selectively using different rendering techniques for different subsets of a 3D data set. Different structures within the data set are rendered locally on an object-by-object basis by either direct volume rendering (DVR), maximum-intensity projection (MIP), surface rendering, value integration (X-ray-like images) or non-photorealistic rendering (NPR). All the results of subsequent object renderings are combined globally in a merging step (usually compositing in our case). This allows us to selectively choose the most suitable technique for depicting each object within the data while keeping the amount of information contained in the image at a reasonable level. This is especially useful when inner structures should be visualized together with semi-transparent outer parts, similar to the focus+context approach known from information visualization. We also present an implementation of our approach which allows us to explore volumetric data using two-level rendering at interactive frame rates     

Real-time Biomechanically-based Muscle Volume Deformation using FEM

This paper presents a voxel-based biomechanical model for muscle deformation using finite element method (FEM) and volume graphics. Hierarchical voxel meshes are reconstructed from filtered segmented muscle images followed by FEM simulation and volume rendering. Physiological muscle force is considered and linear elastic muscle models for both static and dynamic cases are simulated by FEM. Voxel-based wireframe, polygon surface rendering, and volume rendering techniques are applied to show real-time muscle deformation processes as well as realistic animations.     

RenderBots—Multi-Agent Systems for Direct Image Generation

The term stroke‐based rendering collectively describes techniques where images are generated from elements that are usually larger than a pixel. These techniques lend themselves well for rendering artistic styles such as stippling and hatching. This paper presents a novel approach for stroke‐based rendering that exploits multi‐agent systems. RenderBots are individual agents each of which in general represents one stroke. They form a multi‐agent system and undergo a simulation to distribute themselves in the environment. The environment consists of a source image and possibly additional G‐buffers. The final image is created when the simulation is finished by having each RenderBot execute its painting function. RenderBot classes differ in their physical behavior as well as their way of painting so that different styles can be created in a very flexible way.