FLoD: Integrating Flexible Level of Detail into 3D Gaussian Splatting for Customizable Rendering

Yunji Seo,Young Sun Choi,Hyun Seung Son,Youngjung Uh
2024-08-23
Abstract:3D Gaussian Splatting (3DGS) achieves fast and high-quality renderings by using numerous small Gaussians, which leads to significant memory consumption. This reliance on a large number of Gaussians restricts the application of 3DGS-based models on low-cost devices due to memory limitations. However, simply reducing the number of Gaussians to accommodate devices with less memory capacity leads to inferior quality compared to the quality that can be achieved on high-end hardware. To address this lack of scalability, we propose integrating a Flexible Level of Detail (FLoD) to 3DGS, to allow a scene to be rendered at varying levels of detail according to hardware capabilities. While existing 3DGSs with LoD focus on detailed reconstruction, our method provides reconstructions using a small number of Gaussians for reduced memory requirements, and a larger number of Gaussians for greater detail. Experiments demonstrate our various rendering options with tradeoffs between rendering quality and memory usage, thereby allowing real-time rendering across different memory constraints. Furthermore, we show that our method generalizes to different 3DGS frameworks, indicating its potential for integration into future state-of-the-art developments. Project page: <a class="link-external link-https" href="https://3dgs-flod.github.io/flod.github.io/" rel="external noopener nofollow">this https URL</a>
Computer Vision and Pattern Recognition
What problem does this paper attempt to address?
The paper aims to address the issue of high memory consumption in 3D Gaussian Splatting (3DGS) during the rendering process. Specifically: - **Memory Consumption Issue**: 3DGS achieves fast and high-quality rendering by using a large number of small Gaussian splats, which leads to significant memory consumption. This reliance on numerous Gaussian splats limits the application of 3DGS models on low-cost devices, as these devices have limited memory capacity. - **Conflict Between Rendering Quality and Hardware Performance**: Simply reducing the number of Gaussian splats to fit devices with smaller memory would result in a noticeable decline in rendering quality, failing to achieve the effects possible with high-end hardware. To solve the above issues, the authors propose integrating "Flexible Level of Detail" (FLoD) into 3DGS, allowing scenes to be rendered at different levels of detail based on the hardware's capabilities. This approach not only provides customizable rendering options across different hardware configurations but also maintains a balance between rendering quality and memory usage, enabling real-time rendering across varying memory constraints. Additionally, experiments show that this method can be easily integrated into different 3DGS frameworks and is expected to be widely applied in future research.