Recent AI-based video editing has enabled users to edit videos through simple text prompts, significantly simplifying the editing process. However, recent zero-shot video editing techniques primarily focus on global or single-object edits, which can lead to unintended changes in other parts of the video. When multiple objects require localized edits, existing methods face challenges, such as unfaithful editing, editing leakage, and lack of suitable evaluation datasets and metrics. To overcome these limitations, we propose Probability Redistribution for Instance-aware Multi-object Video Editing (PRIMEdit). PRIMEdit is a zero-shot framework that introduces two key modules: (i) Instance-centric Probability Redistribution (IPR) to ensure precise localization and faithful editing and (ii) Disentangled Multi-instance Sampling (DMS) to prevent editing leakage. Additionally, we present our new MIVE Dataset for video editing featuring diverse video scenarios, and introduce the Cross-Instance Accuracy (CIA) Score to evaluate editing leakage in multi-instance video editing tasks. Our extensive qualitative, quantitative, and user study evaluations demonstrate that PRIMEdit significantly outperforms recent state-of-the-art methods in terms of editing faithfulness, accuracy, and leakage prevention, setting a new benchmark for multi-instance video editing.
In this work, we present PRIMEdit, a zero-shot multi-instance video editing framework that disentangles the multi-instance video editing process, achieving faithful edits and reduced attention leakage. Our PRIMEdit effectively disentangles the multi-instance video editing process through our (i) Instance-centric Probability Redistribution (IPR) which enhances editing localization and faithfulness and (ii) Disentangled Multi-instance Sampling (DMS) which reduces editing leakage.
Since our sampling requires that the edited objects appear within their respective masks, we propose our IPR (shown below) to ensure that this condition is consistently met.
In our DMS (shown below), we independently modify each instance using the series noise sampling (yellow box), and the multiple denoised instance latents are harmonized through the parallel noise sampling (blue box) preceded by the latent fusion (green box) and re-inversion (purple box).
We also present our new MIVE Dataset specifically designed for multi-instance video editing tasks. MIVE Dataset features 200 diverse videos sourced from the VIPSeg dataset.
We generated and summarized the source captions using LLaVa and Llama 3, respectively. We then manually inserted tags in the source captions to establish instance-to-mask correspondence. Finally, we generated the target edit captions using Llama 3. We show a sample input video and source and target captions below. The target instance captions are color-coded to match the color of the masks.
Source Caption: In a domestic setting, a person in a gray hoodie stands in front of
washing machine A and washing machine B against a blue wall, with a blue
recycling trash can to the left.
Source Video: |
Target Caption: In a domestic setting, an alien
stands in front of oven
and yellow washing machine against
a blue wall, with a blue recycling trash can to the left.
Masked Source Video: |
---|
Hover over the videos to see the target captions and original video.
Single-Instance Editing: MIVE Dataset Multi-Instance Editing: Video-in-the-Wild |
Multi-Instance Editing: MIVE Dataset Partial Instance Editing: Video-in-the-Wild |
---|
Attention leakage examples are shown in green arrows while unfaithful editing examples are shown in red arrows. The target instance captions are color-coded to match the color of the masks.
Input | Masked Input | PRIMEdit (Ours) | ControlVideo | FLATTEN |
---|---|---|---|---|
FreSCo | TokenFlow | RAVE | Ground-A-Video | VideoGrain |
Input | Masked Input | PRIMEdit (Ours) | ControlVideo | FLATTEN |
---|---|---|---|---|
FreSCo | TokenFlow | RAVE | Ground-A-Video | VideoGrain |
Input | Masked Input | PRIMEdit (Ours) | ControlVideo | FLATTEN |
---|---|---|---|---|
FreSCo | TokenFlow | RAVE | Ground-A-Video | VideoGrain |
Input | Masked Input | PRIMEdit (Ours) | ControlVideo | FLATTEN |
---|---|---|---|---|
FreSCo | TokenFlow | RAVE | Ground-A-Video | VideoGrain |
Input | Masked Input | PRIMEdit (Ours) | ControlVideo | FLATTEN |
---|---|---|---|---|
FreSCo | TokenFlow | RAVE | Ground-A-Video | VideoGrain |
Input | Masked Input | PRIMEdit (Ours) | ControlVideo | FLATTEN |
---|---|---|---|---|
FreSCo | TokenFlow | RAVE | Ground-A-Video | VideoGrain |
@misc{teodoro2025primedit,
title={PRIMEdit: Probability Redistribution for Instance-aware Multi-object Video Editing with Benchmark Dataset},
author={Samuel Teodoro and Agus Gunawan and Soo Ye Kim and Jihyong Oh and Munchurl Kim},
year={2025},
eprint={2412.12877},
archivePrefix={arXiv},
}