VideoCoF: Unified Video Editing with Temporal Reasoner

TL;DR

VideoCoF introduces a novel video editing framework that combines explicit reasoning with diffusion models to achieve precise, mask-free, and length-extrapolatable video editing, validated on a new benchmark.

Contribution

The paper proposes VideoCoF, a Chain-of-Frames approach that enforces reasoning before editing, enabling unified, precise, and efficient video editing without user masks.

Findings

Achieves state-of-the-art results on VideoCoF-Bench.

Requires only 50k video pairs for training.

Enables length extrapolation beyond training duration.

Abstract

Existing video editing methods face a critical trade-off: expert models offer precision but rely on task-specific priors like masks, hindering unification; conversely, unified temporal in-context learning models are mask-free but lack explicit spatial cues, leading to weak instruction-to-region mapping and imprecise localization. To resolve this conflict, we propose VideoCoF, a novel Chain-of-Frames approach inspired by Chain-of-Thought reasoning. VideoCoF enforces a ``see, reason, then edit" procedure by compelling the video diffusion model to first predict reasoning tokens (edit-region latents) before generating the target video tokens. This explicit reasoning step removes the need for user-provided masks while achieving precise instruction-to-region alignment and fine-grained video editing. Furthermore, we introduce a RoPE alignment strategy that leverages these reasoning tokens to ensure motion alignment and enable length extrapolation beyond the training duration. We demonstrate that with a minimal data cost of only 50k video pairs, VideoCoF achieves state-of-the-art performance on VideoCoF-Bench, validating the efficiency and effectiveness of our approach. Our code, weight, data are available at https://github.com/knightyxp/VideoCoF.