Paper 2512.17650
Region-Constraint In-Context Generation for Instructional Video Editing
- Published
- Dec 2025
- Research lab
- Independent
- Citations
- 14
- GitHub
- 174 stars
01 In brief
Summary
ReCo is a novel framework for instruction-based video editing that uses in-context generation with region constraints.
It concatenates source and target videos for joint denoising and introduces two regularization terms: latent-space regularization increases latent discrepancy in editing regions while reducing it in non-editing areas, and attention-space regularization suppresses token interference from the source editing region while strengthening attention to the target background.
The authors also introduce ReCo-Data, a dataset of 500K high-quality instruction-video pairs covering object add, remove, replace, and style transfer tasks.
Experiments on four tasks show ReCo outperforms baselines like Ditto, Lucy-Edit, and InsViE, achieving higher scores in edit accuracy, video naturalness, and video quality.
Ablations confirm the importance of both constraints.
The framework demonstrates strong generalization to creative edits.
02 From the paper
Abstract
The In-context generation paradigm recently has demonstrated strong power in instructional image editing with both data efficiency and synthesis quality. Nevertheless, shaping such in-context learning for instruction-based video editing is not trivial. Without specifying editing regions, the results can suffer from the problem of inaccurate editing regions and the token interference between editing and non-editing areas during denoising. To address these, we present ReCo, a new instructional video editing paradigm that novelly delves into constraint modeling between editing and non-editing regions during in-context generation. Technically, ReCo width-wise concatenates source and target video for joint denoising. To calibrate video diffusion learning, ReCo capitalizes on two regularization terms, i.e., latent and attention regularization, conducting on one-step backward denoised latents and attention maps, respectively. The former increases the latent discrepancy of the editing region between source and target videos while reducing that of non-editing areas, emphasizing the modification on editing area and alleviating outside unexpected content generation. The latter suppresses the attention of tokens in the editing region to the tokens in counterpart of the source video, thereby mitigating their interference during novel object generation in target video. Furthermore, we propose a large-scale, high-quality video editing dataset, i.e., ReCo-Data, comprising 500K instruction-video pairs to benefit model training. Extensive experiments conducted on four major instruction-based video editing tasks demonstrate the superiority of our proposal.