TripleFlow: Training-Free Video Object Removal by Bridging Residual Editing and Native Generation

2026-09-30

AI Summary

This study addresses the failure of background reconstruction and ghosting artifacts in training-free video object removal caused by insufficient generative capacity. To this end, we propose TripleFlow, a tightly coupled multi-stream diffusion architecture comprising source, residual, and synthesis streams that enables synergistic optimization of erasure and generation. Furthermore, we introduce a novel feedback mechanism that continuously injects newly generated backgrounds into the editing trajectory. Combined with residual suppression and native background synthesis algorithms, this approach ensures spatiotemporal consistency without requiring additional training. Extensive experiments demonstrate that TripleFlow establishes new state-of-the-art performance across five benchmarks, significantly outperforming existing methods and substantially improving both reconstruction fidelity and temporal coherence.

Abstract

Video object removal presents a uniquely difficult editing challenge. Because a removal prompt specifies only what to erase rather than what to generate, the model must infer and reconstruct a highly specific occluded background entirely from the surrounding context. Existing training-free methods struggle with this because their editing mechanisms act primarily as localized erasers. They fail to actively synthesize the missing background details and often leave behind ghosting artifacts. To solve this, we propose TripleFlow, a training-free framework that tightly couples erasure and generation. It coordinates a source flow, a residual flow, and a synthesis flow throughout the entire process. By reusing a single target prediction, the residual flow isolates and suppresses the object, while the synthesis flow independently reconstructs the occluded background. Crucially, TripleFlow injects this newly synthesized background back into the editing trajectory at every step. This continuous feedback loop ensures that the generated structures actively guide the removal process, achieving seamless completion that is spatiotemporally consistent with the unedited scene. Extensive evaluations across five challenging benchmarks demonstrate that TripleFlow establishes a new state-of-the-art, significantly outperforming existing baselines in both reconstruction fidelity and temporal consistency.

Problem

• Video Object Removal
• Training-Free
• Background Reconstruction
• Ghosting Artifacts
• Video Editing

Innovation

• Training-Free Video Object Removal
• TripleFlow
• Residual Editing
• Native Generation
• Spatiotemporal Consistency