Real-Time Rendering of Temporal Volumetric Data on a GPU

Lecture Notes in Computer Science, 2007

Volumetric data rendering has become an important tool in various medical procedures as it allows the unbiased visualization of fine details of volumetric medical data (CT, MRI, fMRI). However, due to the large amount of computation involved, the rendering time increases dramatically as the size of the data set grows. This paper presents several acceleration techniques of volume rendering using general-purpose GPU. Some techniques enhance the rendering speed of software ray casting based on voxels' opacity information, while the others improve traditional hardware-accelerated object-order volume rendering. Remarkable speedups are observed using the proposed GPU-based algorithm from experiments on routine medical data sets.

IEEE Transactions on Visualization and Computer Graphics, 2000

Medical volumetric imaging requires high fidelity, high performance rendering algorithms. We motivate and analyze new volumetric rendering algorithms that are suited to modern parallel processing architectures. First, we describe the three major categories of volume rendering algorithms and confirm through an imaging scientist-guided evaluation that ray-casting is the most acceptable. We describe a thread-and data-parallel implementation of ray-casting that makes it amenable to key architectural trends of three modern commodity parallel architectures: multi-core, GPU, and upcoming Intel Larrabee. We achieve more than an order of magnitude performance improvement on a number of large 3D medical datasets. We further describe a data compression scheme that significantly reduces data-transfer overhead. This allows our approach to scale well to large numbers of Larrabee cores.

International Journal of Online and Biomedical Engineering (iJOE)

One of the most valuable medical imaging visualizations or computer-aided diagnosis is Volume rendering (VR). This survey’s objective is reviewing and comparing between several methods and techniques of VR, for a better and more comprehensive reading and learning of both pros and cons of each method, and their use cases.

International Journal of Biomedical Imaging, 2015

Fourier volume rendering (FVR) is a significant visualization technique that has been used widely in digital radiography. As a result of its O(2 log) time complexity, it provides a faster alternative to spatial domain volume rendering algorithms that are O(3) computationally complex. Relying on the Fourier projection-slice theorem, this technique operates on the spectral representation of a 3D volume instead of processing its spatial representation to generate attenuation-only projections that look like X-ray radiographs. Due to the rapid evolution of its underlying architecture, the graphics processing unit (GPU) became an attractive competent platform that can deliver giant computational raw power compared to the central processing unit (CPU) on a per-dollar-basis. The introduction of the compute unified device architecture (CUDA) technology enables embarrassingly-parallel algorithms to run efficiently on CUDA-capable GPU architectures. In this work, a high performance GPU-accelerated implementation of the FVR pipeline on CUDA-enabled GPUs is presented. This proposed implementation can achieve a speed-up of 117x compared to a single-threaded hybrid implementation that uses the CPU and GPU together by taking advantage of executing the rendering pipeline entirely on recent GPU architectures.

2013

Great advancements in commodity graphics hardware have favored GPU-based volume rendering as the main adopted solution for interactive exploration of rectilinear scalar volumes on commodity platforms. Nevertheless, long data transfer times and GPU memory size limitations are often the main limiting factors, especially for massive, time-varying or multi-volume visualization, or for networked visualization on the emerging mobile devices. To address this issue, a variety of level-of-detail data representations and compression techniques have been introduced. In order to improve capabilities and performance over the entire storage, distribution and rendering pipeline, the encoding/decoding process is typically highly asymmetric, and systems should ideally compress at data production time and decompress on demand at rendering time. Compression and level-of-detail pre-computation does not have to adhere to real-time constraints and can be performed off-line for high quality results. In co...

We present a method for fast volume rendering using graphics hardware (GPU). To our knowledge, it is the first implementation on the GPU. Based on the Shear-Warp algorithm, our GPU-based method provides real-time frame rates and outperforms the CPU-based implementation. When the number of slices is not sufficient, we add in-between slices computed by interpolation. This improves then the quality of the rendered images. We have also implemented the ray marching algorithm on the GPU. The results generated by the three algorithms (CPU-based and GPU-based Shear-Warp, GPU-based Ray Marching) for two test models has proved that the ray marching algorithm outperforms the shear-warp methods in terms of speed up and image quality.

Computer Methods and Programs in Biomedicine, 2008

In medical area, interactive three-dimensional volume visualization of large volume datasets is a challenging task. One of the major challenges in graphics processing unit (GPU)-based volume rendering algorithms is the limited size of texture memory imposed by current GPU architecture. We attempt to overcome this limitation by rendering only visible parts of large CT datasets. In this paper, we present an efficient, high-quality volume rendering algorithm using GPUs for rendering large CT datasets at interactive frame rates on standard PC hardware. We subdivide the volume dataset into uniform sized blocks and take advantage of combinations of early ray termination, empty-space skipping and visibility culling to accelerate the whole rendering process and render visible parts of volume data. We have implemented our volume rendering algorithm for a large volume data of 512 × 304 × 1878 dimensions (visible female), and achieved real-time performance (i.e., 3-4 frames per second) on a Pentium 4 2.4 GHz PC equipped with NVIDIA Geforce 6600 graphics card (256 MB video memory). This method can be used as a 3D visualization tool of large CT datasets for doctors or radiologists.

Computer Graphics Forum, 2014

Figure 1: Frames from real-time rendering of animated supernova data set (432 3 × 60, float-18GB), compressed using sparse coding of voxel blocks [GIM12] (block size 6, sparsity 4) at 0.10 bps, PSNR 46.57 dB. The full compressed dataset (184MB) fits in GPU memory and is available for low-latency local and transient decoding during rendering. Data made available by Dr.

2003

For volume rendering of regular grids the display of view-plane aligned slices has proven to yield both good quality and performance. In this paper we demonstrate how to merge the most important extensions of the original 3D slicing approach, namely the pre-integration technique, volumetric clipping, and advanced lighting. Our approach allows the suppression of clipping artifacts and achieves high quality while offering the flexibility to explore volume data sets interactively with arbitrary clip objects. We also outline how to utilize the proposed volumetric clipping approach for the display of segmented data sets. Moreover, we increase the rendering quality by implementing efficient over-sampling with the pixel shader of consumer graphics accelerators. We give prove that at least 4times over-sampling is needed to reconstruct the ray integral with sufficient accuracy even with pre-integration. As an alternative to this brute-force over-sampling approach we propose a hardware-accelerated ray caster which is able to perform over-sampling only where needed and which is able to gain additional speed by early ray termination and space leaping. CR Category: I.3.7 [Computer Graphics]: Three-Dimensional Graphics and Realism.

This paper presents Direct Volume Rendering (DVR) improvement strategies, which provide new opportunities for scientific and medical visualization which are not available in due measure in analogues: 1) multi-volume rendering in a single space of up to 3 volumetric datasets determined in different coordinate systems and having sizes as big as up to 512x512x512 16-bit values; 2) performing the above process in real time on a middle class GPU, e. g. nVidia GeForce GTS 250 512 M B; 3) a custom bounding mesh for more accurate selection of the desired region in addition to the clipping bounding box; 4) simultaneous usage of a number of visualization techniques including the shaded Direct Volume Rendering via the 1D-or 2D-transfer functions, multiple semi-transparent discrete iso-surfaces visualization, M IP, and M IDA. The paper discusses how the new properties affect the implementation of the DVR. In the DVR implementation we use such optimization strategies as the early ray termination and the empty space skipping. The clipping ability is also used as the empty space skipping approach to the rendering performance improvement. We use the random ray start position generation and the further frame accumulation in order to reduce the rendering artifacts. The rendering quality can be also improved by the onthe-fly tri-cubic filtering during the rendering process. Our framework supports 4 different stereoscopic visualization modes. Finally we outline the visualization performance in terms of the frame rates for different visualization techniques on different graphic cards.