# Autoregressive Appearance Prediction for 3D Gaussian Avatars

arXiv: https://arxiv.org/abs/2604.00928 · HTML: https://mlsanity.com/paper/2604.00928/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 40 of its 53 references that are in the map.
- First posted on arXiv 1 Apr 2026.

## Builds on (most cited first, 30 of 40)

- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [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)
- [TensoRF: Tensorial Radiance Fields](https://mlsanity.com/paper/2203.09517/) (2022)
- [Dynamic 3D Gaussians: Tracking by Persistent Dynamic View Synthesis](https://mlsanity.com/paper/2308.09713/) (2023)
- [HumanNeRF: Free-viewpoint Rendering of Moving People from Monocular Video](https://mlsanity.com/paper/2201.04127/) (2022)
- [Animatable Neural Radiance Fields for Modeling Dynamic Human Bodies](https://mlsanity.com/paper/2105.02872/) (2021)
- [gsplat: An Open-Source Library for Gaussian Splatting](https://mlsanity.com/paper/2409.06765/) (2024)
- [NeuMan: Neural Human Radiance Field from a Single Video](https://mlsanity.com/paper/2203.12575/) (2022)
- [3DGS-Avatar: Animatable Avatars via Deformable 3D Gaussian Splatting](https://mlsanity.com/paper/2312.09228/) (2023)
- [GaussianAvatar: Towards Realistic Human Avatar Modeling from a Single Video via Animatable 3D Gaussians](https://mlsanity.com/paper/2312.02134/) (2023)
- [TAVA: Template-free Animatable Volumetric Actors](https://mlsanity.com/paper/2206.08929/) (2022)
- [SplattingAvatar: Realistic Real-Time Human Avatars with Mesh-Embedded Gaussian Splatting](https://mlsanity.com/paper/2403.05087/) (2024)
- [GauHuman: Articulated Gaussian Splatting from Monocular Human Videos](https://mlsanity.com/paper/2312.02973/) (2023)
- [HumanRF: High-Fidelity Neural Radiance Fields for Humans in Motion](https://mlsanity.com/paper/2305.06356/) (2023)
- [H-NeRF: Neural Radiance Fields for Rendering and Temporal Reconstruction of Humans in Motion](https://mlsanity.com/paper/2110.13746/) (2021)
- [HUGS: Human Gaussian Splats](https://mlsanity.com/paper/2311.17910/) (2023)
- [InstantAvatar: Learning Avatars from Monocular Video in 60 Seconds](https://mlsanity.com/paper/2212.10550/) (2022)
- [Real-time Deep Dynamic Characters](https://mlsanity.com/paper/2105.01794/) (2021)
- [Pixel Codec Avatars](https://mlsanity.com/paper/2104.04638/) (2021)
- [GART: Gaussian Articulated Template Models](https://mlsanity.com/paper/2311.16099/) (2023)
- [Drivable 3D Gaussian Avatars](https://mlsanity.com/paper/2311.08581/) (2023)
- [Human Gaussian Splatting: Real-time Rendering of Animatable Avatars](https://mlsanity.com/paper/2311.17113/) (2023)
- [ASH: Animatable Gaussian Splats for Efficient and Photoreal Human Rendering](https://mlsanity.com/paper/2312.05941/) (2023)
- [LiveCap: Real-time Human Performance Capture from Monocular Video](https://mlsanity.com/paper/1810.02648/) (2018)
- [The Power of Points for Modeling Humans in Clothing](https://mlsanity.com/paper/2109.01137/) (2021)
- [AvatarReX: Real-time Expressive Full-body Avatars](https://mlsanity.com/paper/2305.04789/) (2023)
- [Driving-Signal Aware Full-Body Avatars](https://mlsanity.com/paper/2105.10441/) (2021)
- [Modeling Clothing as a Separate Layer for an Animatable Human Avatar](https://mlsanity.com/paper/2106.14879/) (2021)
- [PhysAvatar: Learning the Physics of Dressed 3D Avatars from Visual Observations](https://mlsanity.com/paper/2404.04421/) (2024)

## Abstract (arXiv)

A photorealistic and immersive human avatar experience demands capturing fine, person-specific details such as cloth and hair dynamics, subtle facial expressions, and characteristic motion patterns. Achieving this requires large, high-quality datasets, which often introduce ambiguities and spurious correlations when very similar poses correspond to different appearances. Models that fit these details during training can overfit and produce unstable, abrupt appearance changes for novel poses. We propose a 3D Gaussian Splatting avatar model with a spatial MLP backbone that is conditioned on both pose and an appearance latent. The latent is learned during training by an encoder, yielding a compact representation that improves reconstruction quality and helps disambiguate pose-driven renderings. At driving time, our predictor autoregressively infers the latent, producing temporally smooth appearance evolution and improved stability. Overall, our method delivers a robust and practical path to high-fidelity, stable avatar driving.