Mosaic: Compositional Multi-Concept Erasure via Vector Field Blending

• URL: http://arxiv.org/abs/2605.25574v1
• Date: Mon, 25 May 2026 08:27:26 GMT
• Title: Mosaic: Compositional Multi-Concept Erasure via Vector Field Blending
• Abstract summary: Concept erasure has emerged as a key research direction for ensuring safe and ethical image synthesis in Text-to-Image (T2I) models.<n>We introduce compositional multi-concept erasure, a new task that aims to simultaneously remove multiple target concepts within a single scene.<n>We propose Mosaic, a novel framework for multi-concept erasure in flow-based T2I models, which exploits the spatial locality of target concepts in the vector field.
• Score: 23.575616600476284
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Concept erasure has emerged as a key research direction for ensuring safe and ethical image synthesis in Text-to-Image (T2I) models. While existing studies have explored concept erasure across multiple concepts, they typically assume only a single target concept per image, a limitation increasingly exposed by modern flow-based T2I models, which can generate complex scenes with multiple concepts simultaneously. To address this gap, we introduce compositional multi-concept erasure, a new task that aims to simultaneously remove multiple target concepts within a single scene. We propose CoME-Bench, a benchmark for evaluating compositional multi-concept erasure, which covers both intra- and cross-category scenarios. We further propose Mosaic, a novel framework for multi-concept erasure in flow-based T2I models, which exploits the spatial locality of target concepts in the vector field by dynamically constructing concept-specific masks and selectively blending them without additional optimization. Extensive experiments demonstrate that Mosaic effectively removes multiple target concepts in complex compositional scenes while preserving non-target contexts.

Related papers

• ConceptSplit: Decoupled Multi-Concept Personalization of Diffusion Models via Token-wise Adaptation and Attention Disentanglement [15.939409734710198]
  We present ConceptSplit, a novel framework to split individual concepts through training and inference.<n>Our framework comprises two key components. First, we introduce Token-wise Value Adaptation (ToVA), a merging-free training method.<n>Second, we propose Latent Optimization for Disentangled Attention (LODA), which alleviates attention entanglement during inference.
  arXiv  Detail & Related papers  (2025-10-06T10:22:46Z)
• ICE: Intrinsic Concept Extraction from a Single Image via Diffusion Models [60.30998818833206]
  ICE, short for Intrinsic Concept Extraction, is a novel framework for extracting intrinsic concepts from a single image.<n>Our framework demonstrates superior performance on intrinsic concept extraction from a single image in an unsupervised manner.
  arXiv  Detail & Related papers  (2025-03-25T17:58:29Z)
• Fine-Grained Erasure in Text-to-Image Diffusion-based Foundation Models [56.35484513848296]
  FADE (Fine grained Attenuation for Diffusion Erasure) is an adjacency-aware unlearning algorithm for text-to-image generative models.<n>It removes target concepts with minimal impact on correlated concepts, achieving a 12% improvement in retention performance over state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-03-25T15:49:48Z)
• Modular Customization of Diffusion Models via Blockwise-Parameterized Low-Rank Adaptation [73.16975077770765]
  Modular customization is essential for applications like concept stylization and multi-concept customization.<n>Instant merging methods often cause identity loss and interference of individual merged concepts.<n>We propose BlockLoRA, an instant merging method designed to efficiently combine multiple concepts while accurately preserving individual concepts' identity.
  arXiv  Detail & Related papers  (2025-03-11T16:10:36Z)
• OmniPrism: Learning Disentangled Visual Concept for Image Generation [57.21097864811521]
  Creative visual concept generation often draws inspiration from specific concepts in a reference image to produce relevant outcomes.<n>We propose OmniPrism, a visual concept disentangling approach for creative image generation.<n>Our method learns disentangled concept representations guided by natural language and trains a diffusion model to incorporate these concepts.
  arXiv  Detail & Related papers  (2024-12-16T18:59:52Z)
• AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image Customization [3.5066393042242123]
  We propose AttenCraft, an attention-based method for multiple-concept disentanglement.<n>We introduce an adaptive algorithm based on attention scores to estimate sampling ratios for different concepts.<n>Our model effectively mitigates two issues, achieving state-of-the-art image fidelity and comparable prompt fidelity to baseline models.
  arXiv  Detail & Related papers  (2024-05-28T08:50:14Z)
• MC$^2$: Multi-concept Guidance for Customized Multi-concept Generation [59.00909718832648]
  We propose MC$2$, a novel approach for multi-concept customization.<n>By adaptively refining attention weights between visual and textual tokens, our method ensures that image regions accurately correspond to their associated concepts.<n>Experiments demonstrate that MC$2$ outperforms training-based methods in terms of prompt-reference alignment.
  arXiv  Detail & Related papers  (2024-04-08T07:59:04Z)
• LoRA-Composer: Leveraging Low-Rank Adaptation for Multi-Concept Customization in Training-Free Diffusion Models [33.379758040084894]
  Multi-concept customization emerges as the challenging task within this domain. Existing approaches often rely on training a fusion matrix of multiple Low-Rank Adaptations (LoRAs) to merge various concepts into a single image. LoRA-Composer is a training-free framework designed for seamlessly integrating multiple LoRAs.
  arXiv  Detail & Related papers  (2024-03-18T09:58:52Z)
• Coarse-to-Fine Concept Bottleneck Models [9.910980079138206]
  This work targets ante hoc interpretability, and specifically Concept Bottleneck Models (CBMs) Our goal is to design a framework that admits a highly interpretable decision making process with respect to human understandable concepts, on two levels of granularity. Within this framework, concept information does not solely rely on the similarity between the whole image and general unstructured concepts; instead, we introduce the notion of concept hierarchy to uncover and exploit more granular concept information residing in patch-specific regions of the image scene.
  arXiv  Detail & Related papers  (2023-10-03T14:57:31Z)
• Break-A-Scene: Extracting Multiple Concepts from a Single Image [80.47666266017207]
  We introduce the task of textual scene decomposition. We propose augmenting the input image with masks that indicate the presence of target concepts. We then present a novel two-phase customization process.
  arXiv  Detail & Related papers  (2023-05-25T17:59:04Z)

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.