Pith. sign in

CamGeo: Sparse Camera-Conditioned Image-to-Video Generation with 3D Geometry Priors

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
abstract

Sparse camera-conditioned image-to-video generation presents a pivotal challenge: synthesizing geometrically consistent 3D motion from minimal pose cues. Existing methods, which largely rely on dense supervision or naive interpolation, suffer from severe pose drift and motion discontinuities due to the lack of robust 3D priors. In this paper, we introduce CamGeo, a novel framework that distills rich 3D geometric knowledge from a pre-trained video-to-3D model (VGGT) directly into the diffusion backbone. To achieve this without incurring inference latency, we propose a training-only distillation strategy. Specifically, CamGeo incorporates: (1) keyframe trajectory distillation that enforces cycle-consistency with sparse input poses, (2) cross-frame consistency distillation with both camera trajectory and depth constraints to generate consistent structure across unsupervised frames, and (3) a three-stage coarse-to-fine curriculum learning, progressively scales geometric complexity, from global structure coherence to fine-grained refinement, achieving stable optimization. Extensive experiments demonstrate that CamGeo achieves consistent improvements under various sparsity ratios.

fields

cs.CV 1

years

2026 1

verdicts

CONDITIONAL 1

representative citing papers

MoWorld: A Flash World Model

cs.CV · 2026-07-07 · conditional · novelty 4.0

A 14B-parameter mixture-of-experts world model is distilled to a four-step autoregressive generator and claimed to deliver camera-controllable video at up to 50 FPS on NPUs with 30-50% of prior inference cost.

citing papers explorer

Showing 1 of 1 citing paper.

  • MoWorld: A Flash World Model cs.CV · 2026-07-07 · conditional · none · ref 54 · internal anchor

    A 14B-parameter mixture-of-experts world model is distilled to a four-step autoregressive generator and claimed to deliver camera-controllable video at up to 50 FPS on NPUs with 30-50% of prior inference cost.