See4D: Pose-Free 4D Generation via Auto-Regressive Video Inpainting

TL;DR

See4D introduces a pose-free 4D video generation method that uses virtual cameras and inpainting to synthesize spatiotemporal content without explicit 3D supervision, improving robustness and generalization.

Contribution

The paper proposes a novel pose-free, trajectory-to-camera framework with a view-conditional inpainting model and autoregressive inference, enabling 4D scene synthesis from casual videos without 3D annotations.

Findings

Outperforms pose- and trajectory-conditioned baselines in benchmarks

Achieves superior generalization and scene coherence

Effectively synthesizes 4D content from in-the-wild videos

Abstract

Immersive applications call for synthesizing spatiotemporal 4D content from casual videos without costly 3D supervision. Existing video-to-4D methods typically rely on manually annotated camera poses, which are labor-intensive and brittle for in-the-wild footage. Recent warp-then-inpaint approaches mitigate the need for pose labels by warping input frames along a novel camera trajectory and using an inpainting model to fill missing regions, thereby depicting the 4D scene from diverse viewpoints. However, this trajectory-to-trajectory formulation often entangles camera motion with scene dynamics and complicates both modeling and inference. We introduce See4D, a pose-free, trajectory-to-camera framework that replaces explicit trajectory prediction with rendering to a bank of fixed virtual cameras, thereby separating camera control from scene modeling. A view-conditional video inpainting model is trained to learn a robust geometry prior by denoising realistically synthesized warped images and to inpaint occluded or missing regions across virtual viewpoints, eliminating the need for explicit 3D annotations. Building on this inpainting core, we design a spatiotemporal autoregressive inference pipeline that traverses virtual-camera splines and extends videos with overlapping windows, enabling coherent generation at bounded per-step complexity. We validate See4D on cross-view video generation and sparse reconstruction benchmarks. Across quantitative metrics and qualitative assessments, our method achieves superior generalization and improved performance relative to pose- or trajectory-conditioned baselines, advancing practical 4D world modeling from casual videos.