gaussian-splatting-papers-guide

3D Gaussian Splatting Papers Guide

Overview

3D Gaussian Splatting (3DGS) is a breakthrough technique for real-time radiance field rendering that represents scenes as collections of 3D Gaussians. This curated collection tracks the rapidly evolving 3DGS literature — from the original paper through extensions for dynamic scenes, generation, compression, SLAM, avatars, and more. Essential for researchers in computer vision, graphics, and neural rendering.

Core Paper

@inproceedings{kerbl3Dgaussians,
  title={3D Gaussian Splatting for Real-Time Radiance Field Rendering},
  author={Kerbl, Bernhard and Kopanas, Georgios and Leimk{\"u}hler, Thomas
          and Drettakis, George},
  booktitle={ACM SIGGRAPH 2023},
  year={2023}
}

Key Idea

Input: Multi-view images + SfM point cloud
  ↓
Initialize 3D Gaussians (position, covariance, color, opacity)
  ↓
Differentiable splatting (project Gaussians → image plane)
  ↓
Optimize via photometric loss
  ↓
Adaptive density control (clone, split, prune)
  ↓
Output: Real-time renderable 3D scene (100+ FPS)

Research Landscape

Category Map

Category	Focus	Key Papers
Static Scenes	Quality, compression, anti-aliasing	Mip-Splatting, Compact3D
Dynamic Scenes	Deformable, 4D, temporal	Dynamic3DGS, 4DGS, Deformable3DGS
Generation	Text/image to 3D	DreamGaussian, GaussianDreamer, LGM
SLAM	Real-time mapping	SplaTAM, Gaussian-SLAM, MonoGS
Avatars	Human body/face	GaussianAvatar, HUGS, SplatFace
Autonomous Driving	Street scenes	StreetGaussians, DriveGS
Compression	Storage efficiency	LightGaussian, CompGS
Editing	Scene manipulation	GaussianEditor, GSEditor
Physics	Simulation, deformation	PhysGaussian, Gaussian Splashing
Language	3D understanding	LangSplat, LEGaussians

Tracking New Papers

import requests
from datetime import datetime, timedelta

# Search arXiv for recent 3DGS papers
def search_3dgs_papers(days_back=7):
    """Find recent 3D Gaussian Splatting papers on arXiv."""
    import arxiv

    query = (
        "ti:gaussian splatting OR "
        "abs:3D gaussian splatting OR "
        "abs:3DGS"
    )

    search = arxiv.Search(
        query=query,
        max_results=50,
        sort_by=arxiv.SortCriterion.SubmittedDate,
    )

    cutoff = datetime.now() - timedelta(days=days_back)
    papers = []
    for result in search.results():
        if result.published.replace(tzinfo=None) > cutoff:
            papers.append({
                "title": result.title,
                "authors": [a.name for a in result.authors[:3]],
                "url": result.entry_id,
                "published": result.published.strftime("%Y-%m-%d"),
                "categories": result.categories,
            })
    return papers

recent = search_3dgs_papers(days_back=14)
for p in recent:
    print(f"[{p['published']}] {p['title']}")
    print(f"  {', '.join(p['authors'])} | {p['url']}")