Layered point clouds: a simple and efficient multiresolution structure for distributing and rendering gigantic point-sampled models

We recently introduced an efficient multiresolution structure for distributing and rendering very large point sampled models on consumer graphics platforms [1]. The structure is based on a hierarchy of precomputed object-space point clouds, that are combined coarse-to-fine at rendering time to locally adapt sample densities according to the projected size in the image. The progressive block based refinement nature of the rendering traversal exploits on-board caching and object based rendering APIs, hides out-of-core data access latency through speculative prefetching, and lends itself well to incorporate backface, view frustum, and occlusion culling, as well as compression and view-dependent progressive transmission. The resulting system allows rendering of complex out-of-core models at high frame rates (over 60 M rendered points/second), supports network streaming, and is fundamentally simple to implement. We demonstrate the efficiency of the approach on a number of very large models, stored on local disks or accessed through a consumer level broadband network, including a massive 234 M samples isosurface generated by a compressible turbulence simulation and a 167 M samples model of Michelangelo's St. Matthew. Many of the details of our framework were presented in a previous study. We here provide a more thorough exposition, but also significant new material, including the presentation of a higher quality bottom-up construction method and additional qualitative and quantitative results.     

Efficient image reconstruction for point-based and line-based rendering

We address the problem of an efficient image-space reconstruction of adaptively sampled scenes in the context of point-based and line-based graphics. The image-space reconstruction offers an advantageous time complexity compared to surface splatting techniques and, in fact, our improved GPU implementation performs significantly better than splatting implementations for large point-based models. We discuss the integration of elliptical Gaussian weights for enhanced image quality and generalize the image-space reconstruction to line segments. Furthermore, we present solutions for the efficient combination of points, lines, and polygons in a single image.     

Volume rendering by adaptive refinement

Volume rendering is a technique for visualizing sampled scalar functions of three spatial dimensions by computing 2D projections of a colored semi-transparent gel. This paper presents a volume-rendering algorithm, in which image quality is adaptively refined over time. An initial image is generated by casting a small number of rays into the data, less than one ray per pixel, and interpolating between the resulting colors. Subsequent images are generated by alternately casting more rays and interpolating. The usefulness of these rays is maximized by distributing them according to measures of local image complexity. Examples from two applications are given: molecular graphics and medical imaging.     

Error-controlled real-time cut updates for multi-resolution volume rendering

Multi-resolution techniques are required for rendering large volumetric datasets exceeding the size of the graphics card's memory or even the main memory. The cut through the multi-resolution volume representation is defined by selection criteria based on error metrics. For GPU-based volume rendering, this cut has to fit into the graphics card's memory and needs to be continuously updated due to the interaction with the volume such as changing the area of interest, the transfer function or the viewpoint. We introduce a greedy cut update algorithm based on split-and-collapse operations for updating the cut on a frame-to-frame basis. This approach is guided by a global data-based metric based on the distortion of classified voxel data, and it takes into account a limited download budget for transferring data from main memory into the graphics card to avoid large frame rate variations. Our out-of-core support for handling very large volumes also makes use of split-and-collapse operations to generate an extended cut in the main memory. Finally, we introduce an optimal polynomial-time cut update algorithm, which maximizes the error reduction between consecutive frames. This algorithm is used to verify how close to the optimum our greedy split-and-collapse algorithm performs.     

Ray tracing voxel data via biquadratic local surface interpolation

Various shaded hidden-surface display techniques have been used to render voxel data. In this paper, an approach to using general ray tracing for the rendering of voxel data is presented. Central to this approach is the interpolation of a surface with respect to the volume of a given voxel and its neighbors. Nine columns (of three voxel volumes each) provide sufficient constraints on the integral of a biquadratic function over the column's base to solve its specific coefficients. The use of these locally interpolated surfaces to define a scene to be ray traced is investigated.     

Data structures and the time complexity of ray tracing

The time complexity of ray tracing is a function of the data structures used for space division. Octree and hierarchical extents have been suggested as effective choices. In this paper, complexity parameters are suggested to characterize images and show that both octrees and hierarchies are appropriate choices if given most favorable images. Also, a unified technique is proposed and shown to be better than previous methods for all images. Octrees and hierarchies are particular cases of the new proposed algorithm.     

Heuristics for ray tracing using space subdivision

Ray tracing requires testing of many rays to determine intersections with objects. A way of reducing the computation is to organize objects into hierarchical data structures. We examine two heuristics for space subdivisions using bintrees, one based on the intuition that surface area is a good estimate of intersection probability, one based on the fact that the optimal splitting plane lies between the spatial median and the object median planes of a volume. Traversal algorithms using cross links between nodes are presented as generalizations of ropes in octrees. Simulations of the surface area heuristic and the cross link scheme are presented. These results generalize to other hierarchical data structures.     

Parallel ray tracing based upon a multilevel topological knowledge acquisition of the scene

Including the standard parallelization by grouping primary rays, this paper presents a new parallel ray-timing method based upon a topological knowledge acquisition of the scene. This topological knowledge focuses on relative positions between objects and processes and uses a new type of message. Indeed, instead of exchanging database pages or rays, processes exchange topological information. This information is used by each process to decrease its own list of objects to test against rays The acquisition of information about relative positions between objects and processes is obtained by a careful ordering of he pixel calculation. The processes are dispatched on a computer network including a parallel computer The organization of the processes on this network is a multilevel one leading to different levels of topological message exchanges This method is characterized by topological messages describing the scene, dynamic optimization of the database, easy parallelization on any network (no deadlock, fault tolerance, easily expandable and simple routing), and gives interesting results with true or simulated parallelism.     

真实感图形渲染算法的研究与实现

真实感图形是一种计算机图形生成技术,其关键技术主要包括光照模型、渲染算法、纹理映射、材质设置、颜色量化等方面。本文主要就我们开发的交互式真实感图形渲染系统中采用的各种渲染算法(深度缓冲算法、光线跟踪和辐射度算法)作详细介绍。     

体图形学

体图形学是计算机图形学新崛起的一个重要分支,它在吸收传统计算机图这和计算机影像技术等有关知识的基础上,逐步形成自己独特的概念,理论和方法。     

Parallel rendering of fractal surfaces

Fractal surfaces are a sueful modeling technique for terrain in computer graphics. Although an algorithm exists for ray tracing (Mandelbrot) fractal surfaces, the technique is computationally very expensive. The large degree of parallelism inherent in the problem suggests the use of parallel architectures for generating these images. We describe a parallel rendering algorithm for shared memory MIMD machines which takes advantage of image coherence to reduce computation. This algorithm has, on a Sequent Balance 2100 with 20 processors, demonstrated a near-linear speedup. We examine the possible synchronization bottlenecks by statically assigning different numbers of CPUs to sections of the screen.This work was supported in part by DARPA under contract MDA903-86-C-0182.     

Representation and realistic rendering of natural phenomena with cyclic CSG graphs

A method for ray tracing recursive objects defined by parametric rewriting systems using constructive solid geometry (CSG) as the underlying method of object representation is introduced. Thus, the formal languages of our rewriting systems are subsets of the infinite set of CSG expressions. Instead of deriving such expressions to build up large CSG trees, we translate the systems into cyclic CSG graphs, which can be used directly as an object representation for ray tracing. For this purpose the CSG concept is extended by three new nodes. Selection nodes join all the rules for one grammar symbol, control flow by selecting proper rules, and are end-points of cyclic edges. Transformation nodes map the rays in affine space. Calculation nodes evaluate a finite set of arithmetic expressions to modify global parameters, which effect flow control and transformations. The CSG graphs introduced here are a very compact data structure, much like the describing data set. This property meets our intention to avoid both restrictions of the complexity of the scenes by computer memory and the approximation accuracy of objects.This project was supported by the Fond zur Förderung der wissenschaftlichen Forschung (FWF), Austria, (project no. P09818)     

Two-stage Resampling for Bidirectional Path Tracing with Multiple Light Sub-paths

Recent advances in bidirectional path tracing (BPT) reveal that the use of multiple light sub-paths and the resampling of a small number of these can improve the efficiency of BPT. By increasing the number of pre-sampled light sub-paths, the possibility of generating light paths that provide large contributions can be better explored and this can alleviate the correlation of light paths due to the reuse of pre-sampled light sub-paths by all eye sub-paths. The increased number of pre-sampled light subpaths, however, also incurs a high computational cost. In this paper, we propose a two-stage resampling method for BPT to efficiently handle a large number of pre-sampled light sub-paths. We also derive a weighting function that can treat the changes in path probability due to the two-stage resampling. Our method can handle a two orders of magnitude larger number of presampled light sub-paths than previous methods in equal-time rendering, resulting in stable and better noise reduction than state-of-the-art methods.     

Towards animating ray-traced volume visualization

Ray-tracing volumetric data may take several minutes to compute a single image from a fixed viewpoint. We present techniques that generate approximate ray-traced volumetric images in less than one second per image, after a lengthy initialization process is performed. These approximate images are based on methods that interpolate data sampled at locations on a sphere.     

Exploiting spatial coherence to accelerate radiosity

This paper shows how spatial coherence can be exploited to accelerate form-factor calculation for the radiosity method. A certain number of form factors are calculated by a new technique combining hemisphere and ray tracing, while the other form factors are estimated. This estimation method has reduced form-factor calculation by a factor greate than three, for some test scenes.     

虚拟环境中的三维立体图生成算法

虚拟环境技术是一种高度逼真的模拟人在自然环境中视、听和动等行为的人机界面技术。其中三维立体图生成是一项核心技术。本文对三维立体图显示的原理进行介绍，提出一种立体图生成算法并给出立体图实例。最后对立体图的加速问题进行了讨论。     

Coherent multiresolution isosurface ray tracing

We implement and evaluate a fast ray tracing method for rendering large structured volumes. Input data is losslessly compressed into an octree, enabling residency in CPU main memory. We cast packets of coherent rays through a min/max acceleration structure within the octree, employing a slice-based technique to amortize the higher cost of compressed data access. By employing a multiresolution level of detail (LOD) scheme in conjunction with packets, coherent ray tracing can efficiently render inherently incoherent scenes of complex data. We achieve higher performance with lesser footprint than previous isosurface ray tracers, and deliver large frame buffers, smooth gradient normals and shadows at relatively lesser cost. In this context, we weigh the strengths of coherent ray tracing against those of the conventional single-ray approach, and present a system that visualizes large volumes at full data resolution on commodity computers.