Sparkle: Realizing Lively Instruction-Guided Video Background Replacement via Decoupled Guidance

• URL: http://arxiv.org/abs/2605.06535v1
• Date: Thu, 07 May 2026 16:35:34 GMT
• Title: Sparkle: Realizing Lively Instruction-Guided Video Background Replacement via Decoupled Guidance
• Abstract summary: Background Replacement is central to creative applications such as film production and advertising.<n>Background Replacement requires entirely new, temporally consistent scenes while maintaining accurate foreground-background interactions.<n>We introduce Sparkle, a dataset of 140K video pairs spanning five common background-change themes.
• Score: 54.34473644733862
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In recent years, open-source efforts like Senorita-2M have propelled video editing toward natural language instruction. However, current publicly available datasets predominantly focus on local editing or style transfer, which largely preserve the original scene structure and are easier to scale. In contrast, Background Replacement, a task central to creative applications such as film production and advertising, requires synthesizing entirely new, temporally consistent scenes while maintaining accurate foreground-background interactions, making large-scale data generation significantly more challenging. Consequently, this complex task remains largely underexplored due to a scarcity of high-quality training data. This gap is evident in poorly performing state-of-the-art models, e.g., Kiwi-Edit, because the primary open-source dataset that contains this task, i.e., OpenVE-3M, frequently produces static, unnatural backgrounds. In this paper, we trace this quality degradation to a lack of precise background guidance during data synthesis. Accordingly, we design a scalable pipeline that generates foreground and background guidance in a decoupled manner with strict quality filtering. Building on this pipeline, we introduce Sparkle, a dataset of ~140K video pairs spanning five common background-change themes, alongside Sparkle-Bench, the largest evaluation benchmark tailored for background replacement to date. Experiments demonstrate that our dataset and the model trained on it achieve substantially better performance than all existing baselines on both OpenVE-Bench and Sparkle-Bench. Our proposed dataset, benchmark, and model are fully open-sourced at https://showlab.github.io/Sparkle/.

Related papers

• PAI-Studio: Cinematic Video Background Replacement with Camera-Aware Motion [51.8637101384136]
  We present PAI-Studio, a new reference-conditioned video synthesis task.<n> generating dynamic backgrounds aligned with foreground motion while preserving foreground identity.<n>We further construct a 30K-scale dataset sourced from high-quality films and online videos to support this task.
  arXiv  Detail & Related papers  (2026-05-31T18:45:11Z)
• MoCha:End-to-End Video Character Replacement without Structural Guidance [14.573557179926079]
  MoCha is a framework for replacing video characters with a user-provided identity.<n>We introduce a condition-aware RoPE and employ an RL-based post-training stage.<n>We design three specialized datasets: a high-fidelity rendered dataset built with Unreal Engine 5 (UE5), an expression-driven dataset synthesized by current portrait animation techniques, and an augmented dataset derived from existing video-mask pairs.
  arXiv  Detail & Related papers  (2026-01-13T14:10:34Z)
• Generative Video Matting [57.186684844156595]
  Video matting has traditionally been limited by the lack of high-quality ground-truth data.<n>Most existing video matting datasets provide only human-annotated imperfect alpha and foreground annotations.<n>We introduce a novel video matting approach that can effectively leverage the rich priors from pre-trained video diffusion models.
  arXiv  Detail & Related papers  (2025-08-11T12:18:55Z)
• Video Decomposition Prior: A Methodology to Decompose Videos into Layers [74.36790196133505]
  This paper introduces a novel video decomposition prior VDP' framework which derives inspiration from professional video editing practices.<n>VDP framework decomposes a video sequence into a set of multiple RGB layers and associated opacity levels.<n>We address tasks such as video object segmentation, dehazing, and relighting.
  arXiv  Detail & Related papers  (2024-12-06T10:35:45Z)
• Video DataFlywheel: Resolving the Impossible Data Trinity in Video-Language Understanding [61.89781979702939]
  This study quantitatively reveals an "impossible trinity" among data quantity, diversity, and quality in pre-training datasets. Recent efforts seek to refine large-scale, diverse ASR datasets compromised by low quality through synthetic annotations. We introduce the Video DataFlywheel framework, which iteratively refines video annotations with improved noise control methods.
  arXiv  Detail & Related papers  (2024-09-29T03:33:35Z)
• Transform-Equivariant Consistency Learning for Temporal Sentence Grounding [66.10949751429781]
  We introduce a novel Equivariant Consistency Regulation Learning framework to learn more discriminative representations for each video. Our motivation comes from that the temporal boundary of the query-guided activity should be consistently predicted. In particular, we devise a self-supervised consistency loss module to enhance the completeness and smoothness of the augmented video.
  arXiv  Detail & Related papers  (2023-05-06T19:29:28Z)
• Adaptive Human Matting for Dynamic Videos [62.026375402656754]
  Adaptive Matting for Dynamic Videos, termed AdaM, is a framework for simultaneously differentiating foregrounds from backgrounds. Two interconnected network designs are employed to achieve this goal. We benchmark and study our methods recently introduced datasets, showing that our matting achieves new best-in-class generalizability.
  arXiv  Detail & Related papers  (2023-04-12T17:55:59Z)
• Leaving Reality to Imagination: Robust Classification via Generated Datasets [24.411444438920988]
  Recent research on robustness has revealed significant performance gaps between neural image classifiers trained on datasets similar to the test set. We study the question: How do generated datasets influence the natural robustness of image classifiers? We find that Imagenet classifiers trained on real data augmented with generated data achieve higher accuracy and effective robustness than standard training.
  arXiv  Detail & Related papers  (2023-02-05T22:49:33Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.