IC-Effect: Precise and Efficient Video Effects Editing via In-Context Learning

TL;DR

IC-Effect is a novel framework that enables precise, efficient, and temporally consistent video effects editing using in-context learning and a two-stage training process, even with limited data.

Contribution

It introduces a DiT-based in-context learning approach with spatiotemporal sparse tokenization and a two-stage training strategy for high-quality, controllable video effects editing.

Findings

Achieves seamless effect blending with background preservation.

Demonstrates high fidelity with reduced computation.

Outperforms existing methods in quality and efficiency.

Abstract

We propose \textbf{IC-Effect}, an instruction-guided, DiT-based framework for few-shot video VFX editing that synthesizes complex effects (\eg flames, particles and cartoon characters) while strictly preserving spatial and temporal consistency. Video VFX editing is highly challenging because injected effects must blend seamlessly with the background, the background must remain entirely unchanged, and effect patterns must be learned efficiently from limited paired data. However, existing video editing models fail to satisfy these requirements. IC-Effect leverages the source video as clean contextual conditions, exploiting the contextual learning capability of DiT models to achieve precise background preservation and natural effect injection. A two-stage training strategy, consisting of general editing adaptation followed by effect-specific learning via Effect-LoRA, ensures strong instruction following and robust effect modeling. To further improve efficiency, we introduce spatiotemporal sparse tokenization, enabling high fidelity with substantially reduced computation. We also release a paired VFX editing dataset spanning $15$ high-quality visual styles. Extensive experiments show that IC-Effect delivers high-quality, controllable, and temporally consistent VFX editing, opening new possibilities for video creation.