Paper 2508.13154
4DNeX: Feed-Forward 4D Generative Modeling Made Easy
- Published
- Aug 2025
- Research lab
- Independent
- Citations
- 31
- GitHub
- 840 stars
01 In brief
Summary
4DNeX is the first feed-forward framework for generating 4D (dynamic 3D) scene representations from a single image.
It fine-tunes a pretrained video diffusion model (Wan2.1) to generate a unified 6D video representation (RGB and XYZ sequences) jointly, enabling efficient end-to-end image-to-4D generation.
To address data scarcity, the authors constructed 4DNeX-10M, a large-scale dataset with over 9.2 million video frames and pseudo-4D annotations from sources like DL3DV-10K, RealEstate10K, Pexels, Vimeo, and synthetic data.
They propose simple adaptation strategies including width-wise fusion of RGB and XYZ tokens, XYZ initialization, modality-aware normalization, mask design, and modality-aware token encoding.
The method produces dynamic point clouds that can be rendered into novel-view videos using TrajectoryCrafter.
Experiments show 4DNeX outperforms existing methods in efficiency (15 minutes vs.
60+ minutes) and generalizability, achieving competitive consistency and higher dynamic degree on VBench metrics, with user study preferences favoring it over baselines like Free4D, 4Real, Animate124, and GenXD.
02 From the paper
Abstract
We present 4DNeX, the first feed-forward framework for generating 4D (i.e., dynamic 3D) scene representations from a single image. In contrast to existing methods that rely on computationally intensive optimization or require multi-frame video inputs, 4DNeX enables efficient, end-to-end image-to-4D generation by fine-tuning a pretrained video diffusion model. Specifically, 1) to alleviate the scarcity of 4D data, we construct 4DNeX-10M, a large-scale dataset with high-quality 4D annotations generated using advanced reconstruction approaches. 2) we introduce a unified 6D video representation that jointly models RGB and XYZ sequences, facilitating structured learning of both appearance and geometry. 3) we propose a set of simple yet effective adaptation strategies to repurpose pretrained video diffusion models for 4D modeling. 4DNeX produces high-quality dynamic point clouds that enable novel-view video synthesis. Extensive experiments demonstrate that 4DNeX outperforms existing 4D generation methods in efficiency and generalizability, offering a scalable solution for image-to-4D modeling and laying the foundation for generative 4D world models that simulate dynamic scene evolution.