Summary
Rendering images through cameras involves mapping pixels to bundles of rays, but different camera designs create different ray patterns, making it hard to get consistent, smooth images without glitches called aliasing. The authors found that existing methods often use just one ray per pixel or apply filters in ways that don't fit all camera types, leading to blurry or pixelated results. They created ConeGaussian, a method that better estimates the area each pixel covers by looking at neighboring rays and adjusting the image smoothing accordingly. This approach works across many camera types and improves image quality, especially in fisheye and pinhole cameras, by reducing visual artifacts and maintaining sharpness.
What this means in practice
- •For computer graphics developers: Render sharper and less aliased images from various camera types by integrating ConeGaussian filtering into ray tracing pipelines.
- •For virtual reality content creators: Enhance visual fidelity of VR scenes captured with fisheye or distorted cameras by applying ConeGaussian’s anti-aliasing approach during rendering.
Authors
Deheng Zhang, Letian Shi, Runyi Yang, Zhendong Li, Lei Sun, Kanzhi Wu, Ajad Chhatkuli, Danda Pani Paudel, Luc Van Gool
Abstract
In rendering, a camera is a sampling operator that maps each finite pixel to a bundle of rays. Different camera models change the geometry of this bundle, thus making a unified and faithful rendering formulation challenging. Consequently, Gaussian ray tracing supports generic cameras (with optical center) through their inverse ray mappings, yet typically reduces every pixel to a single center ray. This ignores the camera-dependent pixel footprint, causing aliasing under minification, while unconstrained Gaussians expose unsupported frequencies under magnification. We present ConeGaussian, a camera-model-agnostic anti-aliasing framework for Gaussian ray-based rendering. Instead of defining the pixel filter on a camera-specific image plane, ConeGaussian constructs an anisotropic footprint directly from neighboring rays produced by the camera's native inverse mapping. We derive a closed-form response under a locally linear, depth-local, moment-matched approximation of the finite pixel footprint, while the same geometry defines a per-Gaussian training-frequency floor. Notably, by construction, our filtering principle can be used unmodified across calibrated central camera models and multiple Gaussian ray-rendering backbones. Additionally, unlike in mip-splatting, our scene-space frequency floor and filtering enable trivial composition at render time, allowing us to remove excess blurring. On pinhole and strongly distorted fisheye captures, ConeGaussian consistently improves two distinct ray-based backbones, by up to 4.3 dB at 1/8 resolution, and reduces fisheye LPIPS by 30% where perspective screen-plane footprint formulations are not directly applicable.