Image-based Visualization of Large Volumetric Data Using Moments

Area of Science:

• Computer Graphics
• Scientific Visualization
• Image Processing

Background:

• Visualizing large, complex volumetric datasets interactively is challenging.
• Existing methods often struggle with performance and flexibility for arbitrary data structures.

Purpose of the Study:

• To develop a novel image-based representation for efficient and interactive visualization of volumetric data.
• To enable dynamic transfer function application and modification for enhanced data exploration.

Main Methods:

• A fixed-view image-based representation is created, modeling scalar densities along viewing rays.
• Densities are transformed to the Fourier basis, storing bounded trigonometric moments.
• Adaptive moment determination, novel coding, quantization, and interpolation strategies are employed.

Main Results:

• The representation achieves compact storage and interactive visualization.
• Accurate results are obtained even with view configuration changes.
• The method successfully visualizes large volume datasets and time-dependent SPH data.

Conclusions:

• The proposed image-based representation offers an efficient and flexible solution for interactive volumetric data visualization.
• It enables dynamic exploration and accurate rendering, outperforming existing techniques for complex datasets.