AI Paper Atlas

Research map / 3D vision / Gaussian splatting

Analytic-Splatting: Anti-Aliased 3D Gaussian Splatting via Analytic Integration

Zhihao Liang, Qi Zhang, Wenbo Hu, Ying Feng, Lei ZHU, Kui Jia

ECCV 2024 · 3D vision › Gaussian splatting · approach: Regularization and primitive design for 3DGS · 69 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace point sampling of pixel center with analytic window integral of Gaussians3DGS · Multi-scale Blender SyntheticPSNR avg: 29.77 -> 35.03Improved
replace point sampling with analytic window integral3DGS · Multi-scale Mip-NeRF 360PSNR avg: 27.63 -> 29.51Improved
analytic integral instead of prefiltering (Mip-Splatting)multi-scale Mip-NeRF 360 · Multi-scale Mip-NeRF 360PSNR avg: 29.12 -> 29.51Improved
analytic integral instead of super sampling (3DGS-SS)multi-scale Blender · Multi-scale Blender SyntheticPSNR avg: 32.71 -> 35.03Improved
analytic integral for 2x super-resolution rendering3DGS variants · Multi-scale Mip-NeRF 360, 2x super-resolutionPSNR avg: 26.46 -> 26.90 vs Mip-SplattingImproved
add Mip-Splatting 3D smoothing filter to Analytic-Splatting (ablation)Analytic-Splatting · single-scale Mip-NeRF 360, Tanks&Temples, Deep BlendingPSNR: T&T 23.84 -> 23.91; Deep Blending 29.75 -> 29.71No clear effect
use analytic shading instead of Mip-Splatting for single-scale training and testingMip-Splatting · Mip-NeRF 360 (single-scale)PSNR: 27.57 -> 27.58No clear effect
analytic shading with extra root and exponential operationsAnalytic-Splatting · rendering speedframe rate: 10% lower than Mip-SplattingHurt

Later papers that compared with this method

2 later accepted papers report Analytic-Splatting as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy