Shifting the Breaking Point of Flow Matching for Multi-Instance Editing

• URL: http://arxiv.org/abs/2602.08749v2
• Date: Tue, 10 Feb 2026 12:18:27 GMT
• Title: Shifting the Breaking Point of Flow Matching for Multi-Instance Editing
• Abstract summary: We introduce Instance-Disentangled Attention, a mechanism that partitions joint attention operations and enforces binding between instance-specific textual instructions and spatial regions.<n>Our approach promotes edit disentanglement and locality while preserving global output coherence, enabling single-pass, instance-level editing.
• Score: 47.32746672482526
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Flow matching models have recently emerged as an efficient alternative to diffusion, especially for text-guided image generation and editing, offering faster inference through continuous-time dynamics. However, existing flow-based editors predominantly support global or single-instruction edits and struggle with multi-instance scenarios, where multiple parts of a reference input must be edited independently without semantic interference. We identify this limitation as a consequence of globally conditioned velocity fields and joint attention mechanisms, which entangle concurrent edits. To address this issue, we introduce Instance-Disentangled Attention, a mechanism that partitions joint attention operations, enforcing binding between instance-specific textual instructions and spatial regions during velocity field estimation. We evaluate our approach on both natural image editing and a newly introduced benchmark of text-dense infographics with region-level editing instructions. Experimental results demonstrate that our approach promotes edit disentanglement and locality while preserving global output coherence, enabling single-pass, instance-level editing.

Related papers

• Spatial Temporal Synergy: Balancing Change and Invariance in Text Driven 3D Human Motion Editing [78.44034844227868]
  Text-driven human motion editing aims to modify existing motion sequences according to natural language instructions.<n>Change and Invariance Motion Editing (CIME) is a unified framework that comprehensively decouples change and invariance into spatial pose and temporal rhythm dimensions.<n>Experiments on the MotionFix and STANCE Adjustment datasets demonstrate that CIME achieves state-of-the-art performance in editing alignment and structural fidelity.
  arXiv  Detail & Related papers  (2026-08-17T01:54:44Z)
• Decoupled Guidance: Disentangling Subject and Context Pathways in Text-to-Image Personalization [13.554566178862745]
  Text-to-image personalization aims to generate a user-provided subject in novel scenes described by text.<n>Most existing methods encode subject identity (fidelity) and context (editability) through the same conditioning pathway.<n>We propose Decoupled Guidance (DeGu), a framework that routes subject identity and scene context through two independent guidance streams.
  arXiv  Detail & Related papers  (2026-07-01T10:48:43Z)
• BindEdit: Taming Attention Leakage for Precise Multi-Object Image Editing [8.542586773614113]
  Real image editing enables precise manipulation of visual content, yet existing methods often fail in complex multi-object scenarios.<n>We propose BindEdit, which enforces attention-level constraints within a single diffusion trajectory.<n>BindEdit consistently outperforms existing methods within a single diffusion trajectory.
  arXiv  Detail & Related papers  (2026-06-17T10:32:14Z)
• From Competition to Coopetition: Coopetitive Training-Free Image Editing Based on Text Guidance [11.574335632043491]
  CoEdit is a novel zero-shot framework that transforms attention control from competition to coopetitive negotiation.<n>We present Dual-Entropy Attention Manipulation, which quantifies directional entropic interactions between branches to reformulate attention control as a harmony-maximization problem.<n>We also present Entropic Latent Refinement mechanism to dynamically adjust latent representations over time, minimizing accumulated editing errors.
  arXiv  Detail & Related papers  (2026-04-17T11:10:22Z)
• Layer-wise Instance Binding for Regional and Occlusion Control in Text-to-Image Diffusion Transformers [49.08465459791972]
  Region-instructed layout control in text-to-image generation is highly practical, yet existing methods suffer from limitations.<n>By modeling regional generation as distinct layers and binding them during the generation, we propose LayerBind.
  arXiv  Detail & Related papers  (2026-03-06T00:09:49Z)
• InterCoG: Towards Spatially Precise Image Editing with Interleaved Chain-of-Grounding Reasoning [60.799998743918955]
  We propose a novel text-vision Interleaved Chain-of-Grounding reasoning framework for fine-grained image editing in complex real-world scenes.<n>The key insight of InterCoG is to first perform object position reasoning solely within text.<n>We also propose two auxiliary training modules: multimodal grounding reconstruction supervision and multimodal grounding reasoning alignment.
  arXiv  Detail & Related papers  (2026-03-02T08:13:16Z)
• Rethinking Global Text Conditioning in Diffusion Transformers [28.353061239439587]
  Diffusion transformers typically incorporate textual information via attention layers and a modulation mechanism.<n>Recent approaches discard modulation-based text conditioning and rely exclusively on attention.<n>This paper addresses whether modulation-based text conditioning is necessary and whether it can provide any performance advantage.
  arXiv  Detail & Related papers  (2026-02-09T23:06:58Z)
• ConsistEdit: Highly Consistent and Precise Training-free Visual Editing [17.162316662697965]
  We propose ConsistEdit, a novel attention control method specifically tailored for MM-DiT.<n>It incorporates vision-only attention control, mask-guided pre-attention fusion, and differentiated manipulation of the query, key, and value tokens.<n>It achieves state-of-the-art performance across a wide range of image and video editing tasks, including both structure-consistent and structure-inconsistent scenarios.
  arXiv  Detail & Related papers  (2025-10-20T17:59:52Z)
• FlowDirector: Training-Free Flow Steering for Precise Text-to-Video Editing [2.7248421583285265]
  FlowDirector is a novel inversion-free video editing framework.<n>Our framework models the editing process as a direct evolution in data space.<n>To achieve localized and controllable edits, we introduce an attention-guided masking mechanism.
  arXiv  Detail & Related papers  (2025-06-05T13:54:40Z)
• Tuning-Free Image Editing with Fidelity and Editability via Unified Latent Diffusion Model [60.82962950960996]
  We introduce UnifyEdit, a tuning-free method that performs diffusion latent optimization.<n>We develop two attention-based constraints: a self-attention (SA) preservation constraint for structural fidelity, and a cross-attention (CA) alignment constraint to enhance text alignment.<n>Our approach achieves a robust balance between structure preservation and text alignment across various editing tasks, outperforming other state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-04-08T01:02:50Z)
• Task-Oriented Diffusion Inversion for High-Fidelity Text-based Editing [60.730661748555214]
  We introduce textbfTask-textbfOriented textbfDiffusion textbfInversion (textbfTODInv), a novel framework that inverts and edits real images tailored to specific editing tasks. ToDInv seamlessly integrates inversion and editing through reciprocal optimization, ensuring both high fidelity and precise editability.
  arXiv  Detail & Related papers  (2024-08-23T22:16:34Z)
• Instilling Multi-round Thinking to Text-guided Image Generation [72.2032630115201]
  Single-round generation often overlooks crucial details, particularly in the realm of fine-grained changes like shoes or sleeves. We introduce a new self-supervised regularization, ie, multi-round regularization, which is compatible with existing methods. It builds upon the observation that the modification order generally should not affect the final result.
  arXiv  Detail & Related papers  (2024-01-16T16:19:58Z)
• LIME: Localized Image Editing via Attention Regularization in Diffusion Models [69.33072075580483]
  This paper introduces LIME for localized image editing in diffusion models.<n>LIME does not require user-specified regions of interest (RoI) or additional text input, but rather employs features from pre-trained methods and a straightforward clustering method to obtain precise editing mask.<n>We propose a novel cross-attention regularization technique that penalizes unrelated cross-attention scores in the RoI during the denoising steps, ensuring localized edits.
  arXiv  Detail & Related papers  (2023-12-14T18:59:59Z)
• Textual and Visual Prompt Fusion for Image Editing via Step-Wise Alignment [10.82748329166797]
  We propose a framework that integrates a fusion of generated visual references and text guidance into the semantic latent space of a textitfrozen pre-trained diffusion model.<n>Using only a tiny neural network, our framework provides control over diverse content and attributes, driven intuitively by the text prompt.
  arXiv  Detail & Related papers  (2023-08-30T08:40:15Z)

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.