6 results
Search Results
Now showing 1 - 6 of 6
Item GPU-Based Hyperstreamlines for Diffusion Tensor Imaging(The Eurographics Association, 2006) Reina, G.; Bidmon, K.; Enders, F.; Hastreiter, P.; Ertl, T.; Beatriz Sousa Santos and Thomas Ertl and Ken JoyWe propose a new approach for the visualization of hyperstreamlines, which offers potential for better scalability than the conventional polygon-based approach. Our method circumvents the bandwidth bottleneck between the CPU and GPU by transmitting a small set of parameters for each tube segment and generates the surface directly on the GPU using the classical sphere tracing approach. This reduces the load on the CPU that would otherwise need to provide a suitable level-of-detail representation of the scene, while offering even higher quality in the resulting surfaces since every fragment is traced individually. We demonstrate the effectiveness of this approach by comparing it to the performance and output of conventional visualization tools in the application area of diffusion tensor imaging of human brain MR scans. The method presented here can also be utilized to generate other types of surfaces on the GPU that are too complex to handle with direct ray casting and can therefore be adapted for other applications.Item Hardware-accelerated Extraction and Rendering of Point Set Surfaces(The Eurographics Association, 2006) Tejada, E.; Gois, J. P.; Nonato, L. G.; Castelo, A.; Ertl, T.; Beatriz Sousa Santos and Thomas Ertl and Ken JoyPoint-based models are gaining lately considerable attention as an alternative to traditional surface meshes. In this context, Point Set Surfaces (PSS) were proposed as a modeling and rendering method with important topological and approximation properties. However, ray-tracing PSS is computationally expensive. Therefore, we propose an interactive ray-tracing algorithm for PSS implemented completely on commodity graphics hardware. We also exploit the advantages of PSS to propose a novel technique for extracting surfaces directly from volumetric data. This technique is based on the well known predictor-corrector principle from the numerical methods for solving ordinary differential equations. Our technique provides good approximations to surfaces defined by a certain property in the volume, such as iso-surfaces or surfaces located at regions of high gradient magnitude. Also, local details of the surfaces could be manipulated by changing the local polynomial approximation and the smoothing parameters used. Furthermore, the surfaces generated are smooth and low frequency noise is naturally handled.Item Flexible Interaction with Large Point-Based Datasets(The Eurographics Association, 2006) Rosiuta, A.; Reina, G.; Ertl, T.; Louise M. Lever and Mary McDerbyWe present an application for interacting with large, point-based datasets built with commodity off-the-shelf hardware. Our system works on ordinary workstations with mouse and keyboard interaction as well as on immersive multi-head VR systems with tracked input devices. This approach allows thermodynamics and astrophysics researchers to interactively navigate, filter and inspect large datasets that result from particle-based simulations as used in those areas. We make use of an adaptive, hierarchical rendering approach and extend its data structures to optimize the interaction performance to be usable with datasets of hundreds of thousands of particles. We prove the validity of our concepts with informal user studies both in the application area and the computer science context.Item Pre-integrated Flow Illustration for Tetrahedral Meshes(The Eurographics Association, 2006) Svakhine, N. A.; Ebert, D.; Tejada, E.; Ertl, T.; Gaither, K.; Raghu Machiraju and Torsten MoellerPrevious work has demonstrated the clarity and usefulness of illustrative techniques for visualizing flow data. However, previous systems were limited to applying these techniques to uniform grids. Since unstructured grids have emerged as a common basis for computing flow simulations, we present a method to apply and extend the flow illustration approach to tetrahedral meshes using pre-integrated GPU-accelerated raycasting. Our illustrative rendering techniques can also be applied for other pre-integrated volume rendering systems. Additionally, we explore new feature illustration techniques for flow visualization.Item Optimized Volume Raycasting for Graphics-Hardware-based Cluster Systems(The Eurographics Association, 2006) Müller, C.; Strengert, M.; Ertl, T.; Alan Heirich and Bruno Raffin and Luis Paulo dos SantosIn this paper, we present a sort-last parallel volume rendering system based on single-pass volume raycasting performed in the fragment shader unit. The architecture is aimed for displaying data sets that utilize the total distributed texture memory at interactive framerates. We use uniform texture bricks that are distributed by means of a kd-tree to employ object space partitioning. They are further used for implementing empty-space-skipping and a load balancing mechanism, which also makes use of the kd-tree, to increase the overall performance of the rendering system. Performance numbers are given for a mid-range GPU-cluster system consisting of eight render nodes with an Infiniband interconnection.Item Parallel Texture-Based Vector Field Visualization on Curved Surfaces Using GPU Cluster Computers(The Eurographics Association, 2006) Bachthaler, S.; Strengert, M.; Weiskopf, D.; Ertl, T.; Alan Heirich and Bruno Raffin and Luis Paulo dos SantosWe adopt a technique for texture-based visualization of flow fields on curved surfaces for parallel computation on a GPU cluster. The underlying LIC method relies on image-space calculations and allows the user to visualize a full 3D vector field on arbitrary and changing hypersurfaces. By using parallelization, both the visualization speed and the maximum data set size are scaled with the number of cluster nodes. A sort-first strategy with image-space decomposition is employed to distribute the workload for the LIC computation, while a sort-last approach with an object-space partitioning of the vector field is used to increase the total amount of available GPU memory. We specifically address issues for parallel GPU-based vector field visualization, such as reduced locality of memory accesses caused by particle tracing, dynamic load balancing for changing camera parameters, and the combination of image-space and object-space decomposition in a hybrid approach. Performance measurements document the behavior of our implementation on a GPU cluster with AMD Opteron CPUs, NVIDIA GeForce 6800 Ultra GPUs, and Infiniband network connection.