A Generative Model for Volume Rendering

TL;DR

This paper introduces a GAN-based generative model for volume rendering that enables synthesis, analysis, and editing of volume-rendered images across different rendering techniques and datasets.

Contribution

It presents a novel GAN framework conditioned on viewpoint and transfer functions, facilitating volume analysis and transfer function editing in a view-invariant latent space.

Findings

Effective synthesis of volume-rendered images from various datasets.

Guidance for transfer function editing based on expected image changes.

Compatibility with different rendering methods like ray casting and global illumination.

Abstract

We present a technique to synthesize and analyze volume-rendered images using generative models. We use the Generative Adversarial Network (GAN) framework to compute a model from a large collection of volume renderings, conditioned on (1) viewpoint and (2) transfer functions for opacity and color. Our approach facilitates tasks for volume analysis that are challenging to achieve using existing rendering techniques such as ray casting or texture-based methods. We show how to guide the user in transfer function editing by quantifying expected change in the output image. Additionally, the generative model transforms transfer functions into a view-invariant latent space specifically designed to synthesize volume-rendered images. We use this space directly for rendering, enabling the user to explore the space of volume-rendered images. As our model is independent of the choice of volume rendering process, we show how to analyze volume-rendered images produced by direct and global illumination lighting, for a variety of volume datasets.