# Animatable 3D Gaussian: Fast and High-Quality Reconstruction of Multiple Human Avatars

arXiv: https://arxiv.org/abs/2311.16482 · HTML: https://mlsanity.com/paper/2311.16482/

## Key facts

- Cited by 30 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 14 references that are in the map.
- First posted on arXiv 27 Nov 2023.

## Builds on (most cited first, 7 of 7)

- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022)
- [Neural Sparse Voxel Fields](https://mlsanity.com/paper/2007.11571/) (2020)
- [Geo-PIFu: Geometry and Pixel Aligned Implicit Functions for Single-view Human Reconstruction](https://mlsanity.com/paper/2006.08072/) (2020)
- [Neural Actor: Neural Free-view Synthesis of Human Actors with Pose Control](https://mlsanity.com/paper/2106.02019/) (2021)
- [Generative Neural Articulated Radiance Fields](https://mlsanity.com/paper/2206.14314/) (2022)
- [Capturing and Animation of Body and Clothing from Monocular Video](https://mlsanity.com/paper/2210.01868/) (2022)

## Built on by (most cited first, 30 of 30)

- [3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting](https://mlsanity.com/paper/2312.09228/) (2023)
- [Human Gaussian Splatting: Real-time Rendering of Animatable Avatars](https://mlsanity.com/paper/2311.17113/) (2023)
- [TranSplat: Generalizable 3D Gaussian Splatting from Sparse Multi-View Images with Transformers](https://mlsanity.com/paper/2408.13770/) (2024)
- [LangSplatV2: High-dimensional 3D Language Gaussian Splatting with 450+ FPS](https://mlsanity.com/paper/2507.07136/) (2025)
- [Baking Gaussian Splatting into Diffusion Denoiser for Fast and Scalable Single-stage Image-to-3D Generation and Reconstruction](https://mlsanity.com/paper/2411.14384/) (2024)
- [ZeroHSI: Zero-Shot 4D Human-Scene Interaction by Video Generation](https://mlsanity.com/paper/2412.18600/) (2024)
- [GUAVA: Generalizable Upper Body 3D Gaussian Avatar](https://mlsanity.com/paper/2505.03351/) (2025)
- [GaussianTalker: Real-Time High-Fidelity Talking Head Synthesis with Audio-Driven 3D Gaussian Splatting](https://mlsanity.com/paper/2404.16012/) (2024)
- [Disco4D: Disentangled 4D Human Generation and Animation from a Single Image](https://mlsanity.com/paper/2409.17280/) (2024)
- [LangSurf: Language-Embedded Surface Gaussians for 3D Scene Understanding](https://mlsanity.com/paper/2412.17635/) (2024)
- [A Survey on 3D Human Avatar Modeling -- From Reconstruction to Generation](https://mlsanity.com/paper/2406.04253/) (2024)
- [ToMiE: Towards Explicit Exoskeleton for the Reconstruction of Complicated 3D Human Avatars](https://mlsanity.com/paper/2410.08082/) (2024)
- [Sequential Gaussian Avatars with Hierarchical Motion Context](https://mlsanity.com/paper/2411.16768/) (2024)
- [OccFusion: Rendering Occluded Humans with Generative Diffusion Priors](https://mlsanity.com/paper/2407.00316/) (2024)
- [ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos](https://mlsanity.com/paper/2504.13167/) (2025)
- [Motion-Aware Animatable Gaussian Avatars Deblurring](https://mlsanity.com/paper/2411.16758/) (2024)
- [DAGSM: Disentangled Avatar Generation with GS-enhanced Mesh](https://mlsanity.com/paper/2411.15205/) (2024)
- [RayProNet: A Neural Point Field Framework for Radio Propagation Modeling in 3D Environments](https://mlsanity.com/paper/2406.16907/) (2024)
- [Gaussian Control with Hierarchical Semantic Graphs in 3D Human Recovery](https://mlsanity.com/paper/2405.12477/) (2024)
- [PEAR: Pixel-aligned Expressive humAn mesh Recovery](https://mlsanity.com/paper/2601.22693/) (2026)
- [Deblur-Avatar: Animatable Avatars from Motion-Blurred Monocular Videos](https://mlsanity.com/paper/2501.13335/) (2025)
- [DNF-Avatar: Distilling Neural Fields for Real-time Animatable Avatar Relighting](https://mlsanity.com/paper/2504.10486/) (2025)
- [FastAnimate: Towards Learnable Template Construction and Pose Deformation for Fast 3D Human Avatar Animation](https://mlsanity.com/paper/2512.01444/) (2025)
- [On the Skinning of Gaussian Avatars](https://mlsanity.com/paper/2509.11411/) (2025)
- [HR Human: Modeling Human Avatars with Triangular Mesh and High-Resolution Textures from Videos](https://mlsanity.com/paper/2405.11270/) (2024)
- [FMGS-Avatar: Mesh-Guided 2D Gaussian Splatting with Foundation Model Priors for 3D Monocular Avatar Reconstruction](https://mlsanity.com/paper/2509.14739/) (2025)
- [Towards Efficient 3D Gaussian Human Avatar Compression: A Prior-Guided Framework](https://mlsanity.com/paper/2510.10492/) (2025)
- [InpaintHuman: Reconstructing Occluded Humans with Multi-Scale UV Mapping and Identity-Preserving Diffusion Inpainting](https://mlsanity.com/paper/2601.02098/) (2026)
- [HOIGS: Human-Object Interaction Gaussian Splatting](https://mlsanity.com/paper/2604.04016/) (2026)
- [4DHumanDiff: Direct Text-to-4DGS Generation for Consistent 360-Degree Dynamic Humans](https://mlsanity.com/paper/2607.27634/) (2026)

## Abstract (arXiv)

Neural radiance fields are capable of reconstructing high-quality drivable human avatars but are expensive to train and render and not suitable for multi-human scenes with complex shadows. To reduce consumption, we propose Animatable 3D Gaussian, which learns human avatars from input images and poses. We extend 3D Gaussians to dynamic human scenes by modeling a set of skinned 3D Gaussians and a corresponding skeleton in canonical space and deforming 3D Gaussians to posed space according to the input poses. We introduce a multi-head hash encoder for pose-dependent shape and appearance and a time-dependent ambient occlusion module to achieve high-quality reconstructions in scenes containing complex motions and dynamic shadows. On both novel view synthesis and novel pose synthesis tasks, our method achieves higher reconstruction quality than InstantAvatar with less training time (1/60), less GPU memory (1/4), and faster rendering speed (7x). Our method can be easily extended to multi-human scenes and achieve comparable novel view synthesis results on a scene with ten people in only 25 seconds of training.