AvatarPointillist: AutoRegressive 4D Gaussian Avatarization

TL;DR

AvatarPointillist is a new autoregressive framework that generates dynamic 4D Gaussian avatars from a single image, enabling realistic animation and high-quality rendering.

Contribution

It introduces a decoder-only Transformer for sequential point cloud generation and joint prediction of binding information, advancing avatar creation from minimal input.

Findings

Produces high-quality, photorealistic avatars.

Allows adaptive point density based on subject complexity.

Enables realistic animation through joint prediction.

Abstract

We introduce AvatarPointillist, a novel framework for generating dynamic 4D Gaussian avatars from a single portrait image. At the core of our method is a decoder-only Transformer that autoregressively generates a point cloud for 3D Gaussian Splatting. This sequential approach allows for precise, adaptive construction, dynamically adjusting point density and the total number of points based on the subject's complexity. During point generation, the AR model also jointly predicts per-point binding information, enabling realistic animation. After generation, a dedicated Gaussian decoder converts the points into complete, renderable Gaussian attributes. We demonstrate that conditioning the decoder on the latent features from the AR generator enables effective interaction between stages and markedly improves fidelity. Extensive experiments validate that AvatarPointillist produces high-quality, photorealistic, and controllable avatars. We believe this autoregressive formulation represents a new paradigm for avatar generation, and we will release our code inspire future research.