f i In this paper we present a method for speeding the process of volume rendering a sequence o mages. Speedup is based on exploiting coherency between consecutive images to shorten the n path rays take through the volume. This is achieved by providing each ray with information eeded to leap over the empty space and commence volume traversal at the vicinity of mean- - b ingful data. The algorithm starts by projecting the volume into a C-buffer (Coordinates uffer) which stores, at each pixel location, the object-space coordinates of the first non-empty s t voxel visible from that pixel. For each change in the viewing parameters, the C-buffer i ransformed accordingly. In the case of rotation the transformed C-buffer goes through a pro- - b cess of eliminating coordinates that possibly became hidden. The remaining values in the C uffer serve as an estimate of the point where the new rays should start their volume traverc sal. This space-leaping method can be combined with existing accele...

This paper presents a fast maximum intensity projection technique based on binary shear-warp factorization. The proposed method divides the density domain into a small number of intervals, and to each interval a binary code representation is assigned. In a preprocessing step, an additional volume is created which contains for each voxel the code of the interval enclosing the given voxel density. We present an appropriate data structure for storing this volume and an efficient lookup table technique which can be used to rapidly access a voxel of a certain density code. The volume is efficiently resampled along viewing rays only in voxels where the densities reside in the interval which contains the appropriate maximum value. Keywords: Maximum Intensity Projection, Volume Rendering, Shear-Warp Factorization. 1 INTRODUCTION Maximum Intensity Projection (MIP) is a widely accepted volume rendering technique which can be used especially for the visualization of location, shape, and topolo...

The task of real time rendering of today's volumetric datasets is still being tackled by several research groups. A quick calculation of the amount of computation required for real-time rendering of a high resolution volume puts us in the teraflop range. Yet, the demand to support such rendering capabilities is increasing due to emerging technologies such as virtual surgery simulation and rapid prototyping. There are five main approaches to overcoming this seemingly insurmountable performance barrier: (i) data reduction by means of model extraction or data simplification, (ii) realization of special-purpose volume rendering engines, (iii) software-based algorithm optimization and acceleration, (iv) implementation on general purpose parallel architectures, and (v) use of contemporary of-the-shelf graphics hardware. In this presentation we first describe the vision of real-time high-resolution volume rendering and estimate the computing power it demands. We survey the state-of-the art in...

We consider three-dimensional discrete dynamical system, obtained by the iteration of a noninvertible map of , which simulates the time evolution of an oligopoly game with three competing firms. The model is characterized by the presence of several coexisting stable equilibria, each with its own basin of attraction. In this paper we face the question of the delimitation of the basins and the detection of the global bifurcations that cause the creation of non-connected basins. This requires a study of the global properties of the 3dimensional noninvertible map by the method of critical sets, based on the determination of the contact bifurcations through a systematic computer-assisted study. This requires the visualization of surfaces (the critical surfaces and the basins' boundaries) which sometimes are nested one inside the other. Enhanced graphical methods, based on two-level volume rendering, are employed in order to modulate the opacity of outer objects so that the contacts between the basins' boundaries and critical surfaces can be visualized. This is obtained through the realization of ad-hoc routines, which allow interactive 3D visualization. Key words: Dynamic Games, Discrete Dynamical Systems, noninvertible maps, Computer Graphics, Volume Rendering. 1

. This paper presents a fast volume rotation technique based on binary shear-warp factorization. Unlike many acceleration algorithms this method does not trade image quality for speed and does not require any specialized hardware either. In order to skip precisely the empty regions along the rays to be evaluated a binary volume is generated indicating the locations of the transparent cells. This mask is rotated by an incremental binary shear transformation, executing bitwise boolean operations on integers storing the bits of the binary volume. The ray casting is accelerated using the transformed mask and an appropriate lookup-table technique for finding the first non-transparent cell along each ray. 1 Introduction Direct volume rendering is a flexible but computationally expensive technique for visualization of 3D density arrays. Because of the huge number of voxels to be processed the recent software-only acceleration methods are not fast enough for interactive application...

This paper presents and evaluates a novel method for projecting a three-dimensional volumetric 6-connected object onto a 2D view using seed filling in the view space. Resampling a 6-connected volume by ray casting produces a new 6-connected volume, provided that the sampling rate is high enough. This new volume can be projected by a 3-dimensional seed filling algorithm. We implement and evaluate a volume rendering algorithm which performs the resampling and seed filling in one pass. With this algorithm the projection can be calculated by accessing only part of the voxels within the data set. The method was implemented on a general purpose computer architecture and tested with several artificial and medical volumetric data sets. The frame rate achieved by the algorithm allows nearly interactive (4-5 frames per second) rotation of typical medical volumetric images (MR data 256x256x128 was used) on a common single-processor personal computer with no specialized hardware. The algorithm req...

Introduction Volumetric visualization has been an active research topic since medical devices made possible to acquire volumetric data from inside the human body. In 3D MRI acquisition sequences, the data is collected in the form of an equidistant lattice, with a scalar value associated with each lattice element. The resolution of MRI devices has been constantly increasing, and we have reached the point where the typical amount of voxels (### or greater) is too high for real-time brute-force visualizations on commonly available PC hardware. The apparent usefulness of real-time volumetric visualization has driven the research of this field. Many optimizations, approximations and advanced algorithms have been developed and refined for realtime operation. In the object-order category of algorithms we have e.g. shear-warp[1], splatting[2] and octrees[3]. In the image-order category we have e.g. proximity clouds[4][5] and optimized spacetraversals [6]. In the following, we present a

The task of efficient rendering of today's volumetric datasets is still being tackled by several research groups around the world. A quick calculation of the amount of computation required for interactive rendering of a high resolution volume puts us in the teraflop range. Yet, the demand to support such rendering capabilities is increasing due to emerging technologies such as virtual surgery simulation and rapid prototyping. There are five main approaches to overcoming this seemingly insurmountable performance barrier: (i) data reduction by means of model extraction or data simplification, (ii) realization of special-purpose volume rendering engines, (iii) software-based algorithm optimization and acceleration, (iv) implementation on general purpose parallel architectures, and (v) use of contemporary of-the-shelf graphics hardware. In this presentation we first describe the vision of real-time high-resolution volume rendering and estimate the computing power it demands. W...

this paper we describe a system to show some limited effects on a static toy-car model and present techniques that can be used in similar setups. Our focus is on creating apparent motion for animation

We present a novel technique to accelerate the volume ray casting process used in virtual colonoscopy. Virtual colonoscopy is a less invasive alternative to colonoscopy with potential for wide use in the early detection of colorectal cancer. The idea of our method is to find the exact distance from each image plane pixel to the closest colon wall (boundary) voxel and use this distance information to start ray integration directly from the colon boundary. Our method also improves the rendering time by exploiting the fact that only a small part of the colon is visible from any internal viewing point. We apply a preprocessing visibility determination algorithm to identify the potentially visible part of the colon at each internal view point. The algorithm only projects the potentially visible boundary voxels on to a distance buffer. Using this distance it is possible to skip the space and start the ray casting integration directly from the colon wall. We have used parallel projection and a generic projection template to achieve more acceleration in our implementation. The method has been implemented on an IBM compatible personal computer and tested with a synthetic colon data set.