🤖 AI Summary
Existing 3D Gaussian splatting struggles to model view-dependent reflectance (e.g., specular highlights), leading to artifacts in transparency and reflection rendering. To address this, we propose a view-dependent opacity modeling approach: for the first time, we parameterize opacity as a learnable symmetric matrix coupled with view-direction encoding, enabling physically inspired, view-adaptive Gaussian visibility control—accurately reproducing specular effects while preserving geometric fidelity. Our method is integrated into an extended Gaussian splatting framework supporting efficient differentiable rendering. Evaluated on benchmarks including Mip-NeRF 360 and Tanks & Temples, it achieves state-of-the-art performance, attaining over 60 FPS rendering speed with negligible memory overhead increase.
📝 Abstract
Reconstructing a 3D scene from images is challenging due to the different ways light interacts with surfaces depending on the viewer's position and the surface's material. In classical computer graphics, materials can be classified as diffuse or specular, interacting with light differently. The standard 3D Gaussian Splatting model struggles to represent view-dependent content, since it cannot differentiate an object within the scene from the light interacting with its specular surfaces, which produce highlights or reflections. In this paper, we propose to extend the 3D Gaussian Splatting model by introducing an additional symmetric matrix to enhance the opacity representation of each 3D Gaussian. This improvement allows certain Gaussians to be suppressed based on the viewer's perspective, resulting in a more accurate representation of view-dependent reflections and specular highlights without compromising the scene's integrity. By allowing the opacity to be view dependent, our enhanced model achieves state-of-the-art performance on Mip-Nerf, Tanks&Temples, Deep Blending, and Nerf-Synthetic datasets without a significant loss in rendering speed, achieving>60FPS, and only incurring a minimal increase in memory used.