[2603.02175] Kiwi-Edit: Versatile Video Editing via Instruction and Reference Guidance

Computer Science > Computer Vision and Pattern Recognition arXiv:2603.02175 (cs) [Submitted on 2 Mar 2026] Title:Kiwi-Edit: Versatile Video Editing via Instruction and Reference Guidance Authors:Yiqi Lin, Guoqiang Liang, Ziyun Zeng, Zechen Bai, Yanzhe Chen, Mike Zheng Shou View a PDF of the paper titled Kiwi-Edit: Versatile Video Editing via Instruction and Reference Guidance, by Yiqi Lin and 5 other authors View PDF HTML (experimental) Abstract:Instruction-based video editing has witnessed rapid progress, yet current methods often struggle with precise visual control, as natural language is inherently limited in describing complex visual nuances. Although reference-guided editing offers a robust solution, its potential is currently bottlenecked by the scarcity of high-quality paired training data. To bridge this gap, we introduce a scalable data generation pipeline that transforms existing video editing pairs into high-fidelity training quadruplets, leveraging image generative models to create synthesized reference scaffolds. Using this pipeline, we construct RefVIE, a large-scale dataset tailored for instruction-reference-following tasks, and establish RefVIE-Bench for comprehensive evaluation. Furthermore, we propose a unified editing architecture, Kiwi-Edit, that synergizes learnable queries and latent visual features for reference semantic guidance. Our model achieves significant gains in instruction following and reference fidelity via a progressive multi-stage train...