This paper presents a new 3D RGB image compression scheme designed for interactive real-time applications. In designing our compression method, we have compromised between two important goals: high compression ratio and fast random access ability, and have tried to minimize the overhead caused during runtime reconstruction. Our compression technique is suitable for applications wherein data are accessed in a somewhat unpredictable fashion, and real-time performance of decompression is necessary. The experimental results on three different kinds of 3D images from medical imaging, image-based rendering, and solid texture mapping suggest that the compression method can be used effectively in developing real-time applications that must handle large volume data, made of color samples taken in three- or higher-dimensional space.

Recent advances in realtime image compression and decompression hardware make it possible for a high-performance graphics engine to operate as a rendering server in a networked environment. If the client is a low-end workstation or set-top box, then the rendering task can be split across the two devices. In this paper, we explore one strategy for doing this. For each frame, the server generates a high-quality rendering and a low-quality rendering, subtracts the two, and sends the difference in compressed form. The client generates a matching low quality rendering, adds the decompressed difference image, and displays the composite. Within this paradigm, there is wide latitude to choose what constitutes a high-quality versus low-quality rendering. We have experimented with textured versus untextured surfaces, fine versus coarse tessellation of curved surfaces, Phong versus Gouraud interpolated shading, and antialiased versus nonantialiased edges. In all cases, our polygon-assisted compre...

The classification of volumetric data sets as well as their rendering algorithms are typically based on the representation of the underlying grid. Grid structures based on a Cartesian lattice are the de-facto standard for regular representations of volumetric data. In this paper we introduce a more general concept of regular grids for the representation of volumetric data. We demonstrate that a specific type of regular lattice - the so-called body-centered cubic - is able to represent the same data set as a Cartesian grid to the same accuracy but with 29.3% less samples. This speeds up traditional volume rendering algorithms by the same ratio, which we demonstrate by adopting a splatting implementation for these new lattices. We investigate different filtering methods required for computing the normals on this lattice. The lattice representation results also in lossless compression ratios that are better than previously reported. Although other regular grid structures achieve the same sample efficiency, the body-centered cubic is particularly easy to use. The only assumption necessary is that the underlying volume is isotropic and band-limited - an assumption that is valid for most practical data sets.

We present a simple method for compressing very large and regularly sampled scalar £elds. Our method is particularly attractive when the entire data set does not £t in memory and when the sampling rate is high relative to the feature size of the scalar £eld in all dimensions. Although we report results for R 3 and R 4 data sets, the proposed approach may be applied to higher dimensions. The method is based on the new Lorenzo predictor, introduced here, which estimates the value of the scalar £eld at each sample from the values at processed neighbors. The predicted values are exact when the n-dimensional scalar £eld is an implicit polynomial of degree n−1. Surprisingly, when the residuals (differences between the actual and predicted values) are encoded using arithmetic coding, the proposed method often outperforms wavelet compression in an L ∞ sense. The proposed approach may be used both for lossy and lossless compression and is well suited for out-of-core compression and decompression, because a trivial implementation, which sweeps through the data set reading it once, requires maintaining only a small buffer in core memory, whose size barely exceeds a single (n − 1)-dimensional slice of the data.

Volumetric data sets require enormous storage capacity even at moderate resolution levels. The excessive storage demands not only stress the capacity of the underlying storage and communications systems, but also seriously limit the speed of volume rendering due to data movement and manipulation. A novel volumetric data visualization scheme is proposed and implemented in this work that renders 2D images directly from compressed 3D data sets. The novelty of this algorithm is that rendering is performed on the compressed representation of the volumetric data without pre-decompression. As a result, the overheads associated with both data movement and rendering processing are significantly reduced. The proposed algorithm generalizes previously proposed whole-volume frequency-domain rendering schemes by first dividing the 3D data set into subcubes, transforming each subcube to a frequency-domain representation, and applying the Fourier Projection Theorem to produce the projected 2D images a...

Abstract—Large scale scientific simulation codes typically run on a cluster of CPUs that write/read time steps to/from a single file system. As data sets are constantly growing in size, this increasingly leads to I/O bottlenecks. When the rate at which data is produced exceeds the available I/O bandwidth, the simulation stalls and the CPUs are idle. Data compression can alleviate this problem by using some CPU cycles to reduce the amount of data needed to be transfered. Most compression schemes, however, are designed to operate offline and seek to maximize compression, not throughput. Furthermore, they often require quantizing floating-point values onto a uniform integer grid, which disqualifies their use in applications where exact values must be retained. We propose a simple scheme for lossless, online compression of floating-point data that transparently integrates into the I/O of many applications. A plug-in scheme for data-dependent prediction makes our scheme applicable to a wide variety of data used in visualization, such as unstructured meshes, point sets, images, and voxel grids. We achieve state-of-the-art compression rates and speeds, the latter in part due to an improved entropy coder. We demonstrate that this significantly accelerates I/O throughput in real simulation runs. Unlike previous schemes, our method also adapts well to variable-precision floating-point and integer data. Index Terms—High throughput, lossless compression, file compaction for I/O efficiency, fast entropy coding, range coder, predictive coding, large scale simulation and visualization. 1

Visualization of time-varying volumetric data sets, which may be obtained from numerical simulations or sensing instruments, provides scientists insights into the detailed dynamics of the phenomenon under study. This paper describes our study of a coherent solution based on quantization coupled with octree and difference encoding, and adaptive rendering for efficient visualization of timevarying volumetric data. Quantization is used to attain voxel-level compression and may have a significant influence on the performance of the subsequent encoding and visualization steps. Octree encoding is used for spatial domain compression, and difference encoding for temporal domain compression. In essence, neighboring voxels may be fused into macro voxels if they have similar values, and subtrees at consecutive time steps may be merged if they are identical. The software rendering process is tailored according to the tree structures and the volume visualization process. With the tree representation, selective rendering may be performed very efficiently. Additionally, the I/O costs are reduced. With these combined savings, a higher level of user interactivity is achieved. We have studied a variety of time-varying volume data sets, performed encoding based on data statistics, and optimized the rendering calculations wherever possible. Preliminary tests on workstations have shown in many cases tremendous reduction by as high as 90 % in both storage space and inter-frame delay when compared to direct rendering of the raw data. 1

Visualization of time-varying volumetric data sets, which may be obtained from numerical simulations or sensing instruments, provides scientists insights into the detailed dynamics of the phenomenon under study. This paper describes a coherent solution based on quantization, coupled with octree and difference encoding for visualizing timevarying volumetric data. Quantization is used to attain voxel-level compression and may have a significant influence on the performance of the subsequent encoding and visualization steps. Octree encoding is used for spatial domain compression, and di#erence encoding for temporal domain compression. In essence, neighboring voxels may be fused into macro voxels if they have similar values, and subtrees at consecutive time steps may be merged if they are identical. The software rendering process is tailored according to the tree structures and the volume visualization process. With the tree representation, selective rendering may be performed very efficiently. Additionally, the I/O costs are reduced. With these combined savings, a higher level of user interactivity is achieved. We have studied a variety of timevarying volume datasets, performed encoding based on data statistics, and optimized the rendering calculations wherever possible. Preliminary tests on workstations have shown in many cases tremendous reduction by as high as 90% in both storage space and inter-frame delay.

Gridded volumetric data sets representing simulation or tomography output are commonly visualized by displaying a triangulated isosurface for a particular isovalue. When the grid is stored in a standard format, the entire volume must be loaded from disk, even though only a fraction of the grid cells may intersect the isosurface.

Modern computers are unable to store the complete data of high resolution medical images in main memory. Even on secondary memory (disk), such large datasets are sometimes stored in a compressed form. At rendering time, parts of the volume are requested by the ray tracing algorithm and are loaded from disk. If one is not careful, the same regions may be (decompressed and) loaded to memory several times. Instead, a coherent algorithm should be designed that minimizes this thrashing and optimizes the time and effort spent to (uncompress and) load the volume. We present an algorithm that divides the volume into cubic cells, each (compressed and) stored on disk, in contrast to the more common slice-based storage. At rendering time, each cell is allocated a queue of rays. For a sequence of images, all rays are spawned and queued at the cells they intersect first. Cells are loaded, one at a time, in front-to-back (FTB) order. A loaded cell is rendered by all rays found in its queu...