Tangram-Splatting: Optimizing 3D Gaussian Splatting Through Tangram-inspired Shape Priors

Yi Wang, Ningze Zhong, Minglin Chen, Longguang Wang, Yulan Guo

Abstract: As the growth of VR and AR industry, 3D reconstruction has become a more and more important topic in multimedia. Although 3D Gaussian Splatting is the state-of-the-art method of 3D Reconstruction, it needs a large number of Gaussians to fit a 3D scene due to the Gibbs Phenomenon. The pursuit of compressing 3D Gaussian Splatting and reducing memory overhead has long been a focal point. Embarking on this trajectory, our study delves into this domain, aiming to mitigate these challenges. Inspired by tangram, a Chinese ancient puzzle, we introduce a novel methodology (Tangram-Splatting) that leverages shape priors to optimize 3D scene fitting. Central to our approach is a pioneering technique that diversifies Gaussian function types while preserving algorithmic efficiency. Through exhaustive experimentation, we demonstrate that our method achieves a remarkable average reduction of 62.4\% in memory consumption used to store optimized parameters and decreases the training time by at least 10 minutes, with only marginal sacrifices in PSNR performance, typically under 0.3 dB, and our algorithm is even better on some datasets. This reduction in memory burden is of paramount significance for real-world applications, mitigating the substantial memory footprint and transmission burden traditionally associated with such algorithms. Our algorithm underscores the profound potential of Tangram-Splatting in advancing multimedia applications.

Primary Subject Area: [Experience] Multimedia Applications

Relevance To Conference: Virtual Reality (VR) and Augmented Reality (AR) have always been an important topic of multimedia. In order to foster the industry of VR and AR, 3D reconstruction has become more and more popular, making people more immersed. However, the state-of-the-art algorithm, 3D Gaussian Splatting, has encountered some problems due to its large memory occupation. Our work present a new method which can largely reduce the amount of Gaussians while not influencing the performance much. It has a great potential for future multimedia applications.

Submission Number: 1516