Volumetrically Consistent 3D Gaussian Rasterization
This improves 3D reconstruction and view synthesis for computer vision and graphics applications, offering a more physically accurate alternative to DGS with better performance.
The paper tackles the physical inaccuracies in 3D Gaussian Splatting (3DGS) by volumetrically integrating 3D Gaussians to compute analytic transmittance, resulting in a method that outperforms 3DGS in view synthesis metrics like SSIM and LPIPS and matches state-of-the-art tomography with fewer points.
Recently, 3D Gaussian Splatting (3DGS) has enabled photorealistic view synthesis at high inference speeds. However, its splatting-based rendering model makes several approximations to the rendering equation, reducing physical accuracy. We show that the core approximations in splatting are unnecessary, even within a rasterizer; We instead volumetrically integrate 3D Gaussians directly to compute the transmittance across them analytically. We use this analytic transmittance to derive more physically-accurate alpha values than 3DGS, which can directly be used within their framework. The result is a method that more closely follows the volume rendering equation (similar to ray-tracing) while enjoying the speed benefits of rasterization. Our method represents opaque surfaces with higher accuracy and fewer points than 3DGS. This enables it to outperform 3DGS for view synthesis (measured in SSIM and LPIPS). Being volumetrically consistent also enables our method to work out of the box for tomography. We match the state-of-the-art 3DGS-based tomography method with fewer points. Our code is publicly available at: https://github.com/chinmay0301ucsd/Vol3DGS