dorsal/arxiv
View SchemaLayerGS: Decomposition and Inpainting of Layered 3D Human Avatars via 2D Gaussian Splatting
| Authors | Yinghan Xu, John Dingliana |
|---|---|
| Categories | |
| ArXiv ID | 2601.05853vv1 |
| URL | https://arxiv.org/abs/2601.05853 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
Abstract
We propose a novel framework for decomposing arbitrarily posed humans into animatable multi-layered 3D human avatars, separating the body and garments. Conventional single-layer reconstruction methods lock clothing to one identity, while prior multi-layer approaches struggle with occluded regions. We overcome both limitations by encoding each layer as a set of 2D Gaussians for accurate geometry and photorealistic rendering, and inpainting hidden regions with a pretrained 2D diffusion model via score-distillation sampling (SDS). Our three-stage training strategy first reconstructs the coarse canonical garment via single-layer reconstruction, followed by multi-layer training to jointly recover the inner-layer body and outer-layer garment details. Experiments on two 3D human benchmark datasets (4D-Dress, Thuman2.0) show that our approach achieves better rendering quality and layer decomposition and recomposition than the previous state-of-the-art, enabling realistic virtual try-on under novel viewpoints and poses, and advancing practical creation of high-fidelity 3D human assets for immersive applications. Our code is available at https://github.com/RockyXu66/LayerGS
{
"annotation_id": "91f24bc6-342b-42f9-81db-03f170d5d08d",
"date_created": "2026-02-17T05:53:04.934000Z",
"date_modified": "2026-02-17T05:53:04.934000Z",
"file_hash": "24bd333c44d7c9fc4b0a7b96bd88c9dedcbf4d5b47b9f3ec88b475dd22d56126",
"private": false,
"record": {
"abstract": "We propose a novel framework for decomposing arbitrarily posed humans into animatable multi-layered 3D human avatars, separating the body and garments. Conventional single-layer reconstruction methods lock clothing to one identity, while prior multi-layer approaches struggle with occluded regions. We overcome both limitations by encoding each layer as a set of 2D Gaussians for accurate geometry and photorealistic rendering, and inpainting hidden regions with a pretrained 2D diffusion model via score-distillation sampling (SDS). Our three-stage training strategy first reconstructs the coarse canonical garment via single-layer reconstruction, followed by multi-layer training to jointly recover the inner-layer body and outer-layer garment details. Experiments on two 3D human benchmark datasets (4D-Dress, Thuman2.0) show that our approach achieves better rendering quality and layer decomposition and recomposition than the previous state-of-the-art, enabling realistic virtual try-on under novel viewpoints and poses, and advancing practical creation of high-fidelity 3D human assets for immersive applications. Our code is available at https://github.com/RockyXu66/LayerGS",
"arxiv_id": "2601.05853",
"authors": [
"Yinghan Xu",
"John Dingliana"
],
"categories": [
"cs.CV",
"cs.AI",
"cs.GR"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "LayerGS: Decomposition and Inpainting of Layered 3D Human Avatars via 2D Gaussian Splatting",
"url": "https://arxiv.org/abs/2601.05853",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "51e2652b-9f19-47d3-9835-cb8a95bf243b",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}