Beyond Pixels: From Video Priors to 4D Worlds

Hugging Face Daily Papers Papers

Summary

This paper introduces Latent-to-4D, a method for direct 4D scene generation from video diffusion latents without retraining across generators, achieving better geometry and temporal stability than cascaded approaches.

4D generation synthesizes dynamic 3D scenes from conditions such as text or images. Existing methods either reconstruct generated RGB videos with a separate 4D model or adapt a particular video generator to predict geometry directly. The former suffers from distribution mismatch and error propagation, whereas the latter ties 4D prediction to a specific generator and may require retraining when the generator or conditioning regime changes. We ask whether the final denoised latents of video models that share a variational autoencoder (VAE) can instead provide a reusable interface to explicit 4D prediction. Building on this insight, we introduce direct latent-to-4D generation and instantiate it as Latent-to-4D, which bypasses RGB by aligning a video latent with the token grid of a pretrained 4D decoder and refining it through frame-wise and global spatiotemporal attention. Trained on roughly 1K existing reconstruction clips, a single checkpoint transfers unchanged across multiple video diffusion transformers within the same VAE family. On Text4D-200 and I4D-200, Latent-to-4D surpasses matched same-latent Wan+4RC cascades in projection-based DINO-F1 by 2.88--3.45 and 5.81 points, respectively, while also being preferred by human raters for geometry, temporal stability, and overall quality.
Original Article
View Cached Full Text

Cached at: 08/12/26, 08:21 AM

Paper page - Beyond Pixels: From Video Priors to 4D Worlds

Source: https://huggingface.co/papers/2608.10744

Abstract

Latent-to-4D enables reusable direct 4D generation from video diffusion latents via alignment with a pretrained decoder and spatiotemporal attention, transferring across generators without retraining.

4D generationsynthesizes dynamic 3D scenes from conditions such as text or images. Existing methods either reconstruct generated RGB videos with a separate 4D model or adapt a particular video generator to predict geometry directly. The former suffers from distribution mismatch and error propagation, whereas the latter ties 4D prediction to a specific generator and may require retraining when the generator or conditioning regime changes. We ask whether the final denoised latents of video models that share avariational autoencoder(VAE) can instead provide a reusable interface to explicit 4D prediction. Building on this insight, we introduce directlatent-to-4Dgeneration and instantiate it asLatent-to-4D, which bypasses RGB by aligning a video latent with the token grid of a pretrained4D decoderand refining it through frame-wise and globalspatiotemporal attention. Trained on roughly 1K existing reconstruction clips, a single checkpoint transfers unchanged across multiplevideo diffusion transformerswithin the same VAE family. On Text4D-200 and I4D-200,Latent-to-4Dsurpasses matched same-latent Wan+4RC cascades in projection-based DINO-F1 by 2.88--3.45 and 5.81 points, respectively, while also being preferred by human raters for geometry, temporal stability, and overall quality.

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2608\.10744

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/2608.10744 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.10744 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.10744 in a Space README.md to link it from this page.

Collections including this paper1

Similar Articles

Latent Spatial Memory for Video World Models

Hugging Face Daily Papers

This paper introduces latent spatial memory for video world models, storing 3D scene information directly in diffusion latent space to avoid costly pixel-space reconstruction. The proposed Mirage framework achieves up to 10.57x faster generation and 55x memory reduction while achieving state-of-the-art performance on WorldScore and RealEstate10K.

L2P: Unlocking Latent Potential for Pixel Generation

Hugging Face Daily Papers

The L2P paper introduces a Latent-to-Pixel transfer paradigm that leverages pre-trained latent diffusion models to create efficient pixel-space models capable of 4K generation with minimal training overhead.

4DAnyone: Create Anyone in 4D from a Casual Monocular Video

Hugging Face Daily Papers

4DAnyone reconstructs 4D humans from monocular video by generating multiview-consistent videos and lifting them into 4D Gaussian Splatting, using reference and target context designs to address scaling bottlenecks.