Mip-Splatting: Alias-free 3D Gaussian Splatting

TL;DR

This paper introduces Mip-Splatting, a method that reduces aliasing artifacts in 3D Gaussian Splatting by applying 3D smoothing and Mip filters, improving view synthesis quality across different scales.

Contribution

It proposes a novel 3D smoothing filter and replaces 2D dilation with a Mip filter to eliminate high-frequency artifacts and aliasing in 3D Gaussian Splatting.

Findings

Reduces artifacts when changing sampling rates or zooming.

Improves synthesis quality across multiple scales.

Validates effectiveness through comprehensive evaluation.

Abstract

Recently, 3D Gaussian Splatting has demonstrated impressive novel view synthesis results, reaching high fidelity and efficiency. However, strong artifacts can be observed when changing the sampling rate, \eg, by changing focal length or camera distance. We find that the source for this phenomenon can be attributed to the lack of 3D frequency constraints and the usage of a 2D dilation filter. To address this problem, we introduce a 3D smoothing filter which constrains the size of the 3D Gaussian primitives based on the maximal sampling frequency induced by the input views, eliminating high-frequency artifacts when zooming in. Moreover, replacing 2D dilation with a 2D Mip filter, which simulates a 2D box filter, effectively mitigates aliasing and dilation issues. Our evaluation, including scenarios such a training on single-scale images and testing on multiple scales, validates the effectiveness of our approach.