AvatarPointillist: AutoRegressive 4D Gaussian Avatarization
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866918456253743104 |
|---|---|
| author | Liu, Hongyu Wang, Xuan Wu, Zijian Wang, Yating Wan, Ziyu Ma, Yue Liu, Runtao Zhou, Boyao Shen, Yujun Chen, Qifeng |
| author_facet | Liu, Hongyu Wang, Xuan Wu, Zijian Wang, Yating Wan, Ziyu Ma, Yue Liu, Runtao Zhou, Boyao Shen, Yujun Chen, Qifeng |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_04787 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AvatarPointillist: AutoRegressive 4D Gaussian Avatarization Liu, Hongyu Wang, Xuan Wu, Zijian Wang, Yating Wan, Ziyu Ma, Yue Liu, Runtao Zhou, Boyao Shen, Yujun Chen, Qifeng Computer Vision and Pattern Recognition 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. |
| title | AvatarPointillist: AutoRegressive 4D Gaussian Avatarization |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.04787 |