Multi-view Consistent 3D Gaussian Head Avatars 'without' Multi-view Generation

Hugging Face Daily Papers Papers
3d-gaussian head-avatar multi-view-consistency state-space-model mamba single-shot 3d-generation

Summary

MVCHead is a novel method for generating 3D Gaussian head avatars from single 2D images without multi-view data, using hierarchical state space models and multi-view consistency enforcement.

High-fidelity 3D Gaussian head avatar generation is critical for applications such as AR/VR, telepresence, and digital humans. Existing methods depend on multi-view datasets, 3D captures, or intermediate 2D view synthesis. In contrast, we learn both conditional and unconditional 3D head models from randomly sampled 2D images alone, without using multi-view data, 3D supervision, or intermediate view generation. We introduce MVCHead, a single-shot state space model that enforces multi-view consistency (MVC) directly in the 3D representation while regressing 3D Gaussians under these constraints. At its core, we propose a Hierarchical State Space (HiSS) block that progressively refines Gaussians from coarse to fine, while capturing long-range dependencies. Within each HiSS block, we modify Mamba's standard unidirectional scan with the proposed Hierarchical Bi-directional State Scan (HiBiSS) that aligns recurrence with the axes along which multi-view inconsistencies are strongest. Finally, we design an SE(3) Multi-view Critic that judges whether a set of self-renders arises from a single underlying 3D configuration, rewarding cross-view pixel alignment without observing real multi-view pairs. MVCHead achieves state-of-the-art perceptual quality, surpasses prior methods in both texture and geometric consistency, and maintains comparable shape consistency. To demonstrate scalability, we release FaceGS-10K, the first large-scale dataset of ready-to-use 3D Gaussian head assets for training and evaluation of 3D head models. Project Page and code: https://humansensinglab.github.io/MVCHead/
Original Article
View Cached Full Text

Cached at: 05/29/26, 03:02 PM

Paper page - Multi-view Consistent 3D Gaussian Head Avatars ‘without’ Multi-view Generation

Source: https://huggingface.co/papers/2605.25220

Abstract

A novel single-shot 3D Gaussian head avatar generation method called MVCHead uses hierarchical state space models and multi-view consistency enforcement to create high-fidelity 3D assets from 2D images without requiring multi-view data or 3D supervision.

High-fidelity3D Gaussian head avatargeneration is critical for applications such as AR/VR, telepresence, and digital humans. Existing methods depend onmulti-view datasets, 3D captures, or intermediate 2D view synthesis. In contrast, we learn both conditional and unconditional3D head modelsfrom randomly sampled2D imagesalone, without usingmulti-view data,3D supervision, or intermediate view generation. We introduce MVCHead, a single-shotstate space modelthat enforcesmulti-view consistency(MVC) directly in the3D representationwhile regressing3D Gaussians under these constraints. At its core, we propose aHierarchical State Space(HiSS) block that progressively refines Gaussians from coarse to fine, while capturing long-range dependencies. Within eachHiSS block, we modifyMamba’s standard unidirectional scan with the proposedHierarchical Bi-directional State Scan(HiBiSS) that aligns recurrence with the axes along which multi-view inconsistencies are strongest. Finally, we design anSE(3) Multi-view Criticthat judges whether a set ofself-rendersarises from a single underlying 3D configuration, rewarding cross-view pixel alignment without observing real multi-view pairs. MVCHead achieves state-of-the-art perceptual quality, surpasses prior methods in both texture and geometric consistency, and maintains comparable shape consistency. To demonstrate scalability, we release FaceGS-10K, the first large-scale dataset of ready-to-use3D Gaussianhead assets for training and evaluation of3D head models. Project Page and code: https://humansensinglab.github.io/MVCHead/

View arXiv pageView PDFProject pageGitHubAdd to collection

Get this paper in your agent:

hf papers read 2605\.25220

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2605.25220 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2605.25220 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.25220 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis

Hugging Face Daily Papers

NAMVIS proposes a diffusion-free framework that reformulates sparse-view multi-view image synthesis as geometry-conditioned next-scale autoregression, using Multi-scale Projective Pose Encoding to anchor generation to camera geometry. It outperforms diffusion baselines on Objaverse, GSO, and OmniObject3D in PSNR/SSIM/LPIPS while running over 3x faster.