HPMC is a small OpenGL/C/C++-library that extracts iso-surfaces of volumetric data directly on the GPU.

The library analyzes a lattice of scalar values describing a scalar field that is either stored in a Texture3D or can be accessed through an application-provided snippet of shader code. The output is a sequence of vertex positions and normals that form a triangulation of the iso-surface. HPMC provides traversal code to be included in an application vertex shader, which allows direct extraction in the vertex shader. Using the OpenGL transform feedback mechanism, the triangulation can be stored directly into a buffer object.

(C. Dyken, G. Ziegler, C. Theobalt, H.-P. Seidel, High-speed Marching Cubes using Histogram Pyramids, Computer Graphics Forum 27 (8), 2008.)

This paper by Jansen et al. describes how to utilize current commodity graphics hardware to perform Fourier volume rendering directly on the GPU. The paper presents a novel implementation of the Fast Fourier Transform: This Split-Stream-FFT maps the recursive structure of the FFT to the GPU in an efficient way. Additionally, high-quality resampling within the frequency domain is discussed. The implementation enables visualization of large volumetric data sets at interactive frame rates on a mid-range computer system. (Fourier Volume Rendering on the GPU Using a Split-Stream FFT)

This paper from the VIS Group Stuttgart shows the first volume filtering algorithm that uses OpenGL for the convolution process. Filtering volume data is useful for noise reduction, feature detection, and segmentation. The process is significantly accelerated on SGI graphics workstations with hardware support for two-dimensional image convolution in the frame buffer. Generic 3D convolution can be added as a powerful tool in interactive volume visualization toolkits. See also the project page for more about hardware-based filtering. (Accelerating 3D Convolution using Graphics Hardware. Matthias Hopf and Thomas Ertl. Proc. Visualization 1999, pp 471–474.)

This TVCG paper is an extended version of a Vis 2003 paper, with significantly more detail about the time-dependent, sparse-grid GPU computation strategy used in the level-set solver. The paper describes a 3D-to-2D virtual memory address scheme for packing the narrow-band data into GPU memory. It also adds detail about the GPU-based distance transform computation and the GPU-to-CPU message passing approach. Lastly, the paper describes a volume rendering algorithm for rendering compressed data that provides on-the-fly reconstruction, full trilinear interpolation, and the ability to render from any viewpoint without data duplication. (A Streaming Narrow Band Algorithm: Interactive Computation and Visualization of Level Sets. A. Lefohn, J. Kniss, C. Hansen, and R. Whitaker. Transactions on Visualization and Computer Graphics.)

This IEEE Visualization 2003 tutorial presentation by Aaron Lefohn gives a high-level overview of dynamic volume computation and visualization on GPUs. The talk is part of the tutorial Interactive Visualization of Volumetric Data on Consumer PC Hardware. The first half of the presentation discusses various memory layout options for dynamic volume computation, and the implications of each option on computation and rendering. The second half discusses optimizations and load balancing between the various computational resources: CPU, vertex processor, rasterizer, and fragment processor. (Dynamic Volume
Computation and Visualization on the GPU, by Aaron Lefohn)

This Vis03 paper by Krüger and Westermann addresses the integration of early ray termination and empty-space skipping into texture based volume rendering on graphical processing units (GPU). Therefore, volume ray-casting on programmable graphics hardware is described as an alternative to object-order approaches. The early z-test is exploited to terminate fragment processing once sufficient opacity has been accumulated, and to skip empty space along the rays of sight. Performance gains up to a factor of 3 for typical renditions of volumetric data sets on the ATI 9700 graphics card are demonstrated. (Acceleration Techniques for GPU-based Volume Rendering To appear in IEEE Visualization 2003)

This IEEE Visualization 2003 paper by Sherbondy et al. describes a fast approach to segmentation with simultaneous visualization for medical imaging using the GPU. Specifically, their approach iterates a seeded region growing algorithm, where the seeds are dispersed according to a nonlinear diffusion process. An implicit segmentation surface is then defined according to the population of the seed volume. This paper also introduces the technique of computational masking, whereby the depth buffer is used to limit computation to an active subvolume. This segmentation algorithm is shown to be 10-20 times faster than an SSE2 optimized CPU version of the algorithm. (Fast Volume Segmentation With Simultaneous Visualization Using Programmable Graphics Hardware. Anthony Sherbondy, Mike Houston, Sandy Napel. To appear in IEEE Visualization 2003)

This IEEE Visualization 2003 paper by Lefohn et al. describes the computation and visualization of deforming level-set surfaces on a GPU. The paper describes a new technique for efficiently processing the dynamic, sparse volumes encountered in level-set computations. The technique stores a packed representation of the sparse volume on the GPU and performs both the level-set computation and real-time volume rendering on this packed format. The new level-set solver is shown to be 10-15 times faster than an optimized CPU implementation. The paper demonstrates the new algorithm with a volume segmentation application. (Interactive Deformation and Visualization of Level Set Surfaces using Graphics Hardware. Aaron E. Lefohn, Joe M. Kniss, Charles D. Hansen and Ross T. Whitaker. To appear in IEEE Visualization 2003.)

This Medical Image Computing and Computer Assisted Intervention (MICCAI) 2003 paper by Lefohn et al. describes a brain tumor segmentation study performed with a new GPU-based level-set solver. This paper demonstrates that the ability to interact with a level-set computation in real time enables users to quickly produce segmentations from MRI data that qualitatively and quantitatively compare favorably with expert hand-segmentations. (Interactive, GPU-Based Level Sets for 3D Brain Tumor Segmentation. Aaron E. Lefohn, Joshua E. Cates and Ross T. Whitaker. To appear at “Medical Image Computing and Computer Assisted Intervention,” (MICCAI) 2003.)

This paper by Lefohn et. al. at the University of Utah demonstrates a 3D level-set PDE solver implemented entirely on an ATI Radeon 8500 GPU. It shows that in addition to the basic level-set equations, the second-order mean curvature speed term can be successfully evaluated with only 8-bit textures. The paper demonstrates the solver segmenting the cortical surface from a 256 x 256 x 175 MRI volume and compares the results to a CPU implementation. The object oriented framework with which the solver was built, “Glift,” is also discussed. (A GPU-Based, Three-Dimensional Level Set Solver with Curvature Flow. Aaron Lefohn and Ross Whitaker. University of Utah tech report UUCS-02-017, December, 2002.)