MultiBind: A Benchmark for Attribute Misbinding in Multi-Subject Generation

• URL: http://arxiv.org/abs/2603.21937v1
• Date: Mon, 23 Mar 2026 12:54:36 GMT
• Title: MultiBind: A Benchmark for Attribute Misbinding in Multi-Subject Generation
• Abstract summary: We introduce MultiBind, a benchmark built from real multi-person photographs.<n>We show that MultiBind reveals binding failures that conventional reconstruction metrics miss.
• Score: 30.628326435554182
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Subject-driven image generation is increasingly expected to support fine-grained control over multiple entities within a single image. In multi-reference workflows, users may provide several subject images, a background reference, and long, entity-indexed prompts to control multiple people within one scene. In this setting, a key failure mode is cross-subject attribute misbinding: attributes are preserved, edited, or transferred to the wrong subject. Existing benchmarks and metrics largely emphasize holistic fidelity or per-subject self-similarity, making such failures hard to diagnose. We introduce MultiBind, a benchmark built from real multi-person photographs. Each instance provides slot-ordered subject crops with masks and bounding boxes, canonicalized subject references, an inpainted background reference, and a dense entity-indexed prompt derived from structured annotations. We also propose a dimension-wise confusion evaluation protocol that matches generated subjects to ground-truth slots and measures slot-to-slot similarity using specialists for face identity, appearance, pose, and expression. By subtracting the corresponding ground-truth similarity matrices, our method separates self-degradation from true cross-subject interference and exposes interpretable failure patterns such as drift, swap, dominance, and blending. Experiments on modern multi-reference generators show that MultiBind reveals binding failures that conventional reconstruction metrics miss.

Related papers

• MultiCompose: Multi-Concept Personalized Composition with Per-Subject Attribute Binding [9.605365111072745]
  MultiCompose is a composition framework that decouples per-concept personalization from multi-subject inference.<n>MSP-Bench is a benchmark that jointly evaluates identity fidelity (ID), attribute binding accuracy (BIND), and attribute misalignment (MIS)<n>Experiments show that MultiCompose outperforms existing methods on both conventional metrics and MSP-Bench.
  arXiv  Detail & Related papers  (2026-08-04T14:13:18Z)
• UniCustom: Unified Visual Conditioning for Multi-Reference Image Generation [65.53694602893042]
  We propose a unified visual conditioning framework that fuses ViT and VAE features before VLM encoding.<n>Experiments on two multi-reference generation benchmarks demonstrate that UniCustom consistently improves subject consistency, instruction following, and compositional fidelity.
  arXiv  Detail & Related papers  (2026-05-12T13:10:05Z)
• Towards Generalized Multi-Image Editing for Unified Multimodal Models [56.620038824933566]
  Unified Multimodal Models (UMMs) integrate multimodal understanding and generation.<n>UMMs are limited to maintaining visual consistency and disambiguating visual cues when referencing details across multiple input images.<n>We propose a scalable multi-image editing framework for UMMs that explicitly distinguishes image identities and generalizes to variable input counts.
  arXiv  Detail & Related papers  (2026-01-09T06:42:49Z)
• Towards Unified Semantic and Controllable Image Fusion: A Diffusion Transformer Approach [99.80480649258557]
  DiTFuse is an instruction-driven framework that performs semantics-aware fusion within a single model.<n>Experiments on public IVIF, MFF, and MEF benchmarks confirm superior quantitative and qualitative performance, sharper textures, and better semantic retention.
  arXiv  Detail & Related papers  (2025-12-08T05:04:54Z)
• MultiBanana: A Challenging Benchmark for Multi-Reference Text-to-Image Generation [38.21587139161503]
  $textbfMultiBanana$ is designed to assess the edge of model capabilities by widely covering multi-reference-specific problems at scale.<n>Our analysis reveals their superior performances, typical failure modes, and areas for improvement.<n>MultiBanana will be released as an open benchmark to push the boundaries and establish a standardized basis for fair comparison in multi-reference image generation.
  arXiv  Detail & Related papers  (2025-11-28T08:49:55Z)
• Canvas-to-Image: Compositional Image Generation with Multimodal Controls [51.44122945214702]
  We introduce Canvas-to-Image, a unified framework that consolidates heterogeneous controls into a single canvas interface.<n>Our key idea is to encode diverse control signals into a single composite canvas image that the model can interpret for integrated visual-spatial reasoning.
  arXiv  Detail & Related papers  (2025-11-26T18:59:56Z)
• WithAnyone: Towards Controllable and ID Consistent Image Generation [83.55786496542062]
  Identity-consistent generation has become an important focus in text-to-image research.<n>We develop a large-scale paired dataset tailored for multi-person scenarios.<n>We propose a novel training paradigm with a contrastive identity loss that leverages paired data to balance fidelity with diversity.
  arXiv  Detail & Related papers  (2025-10-16T17:59:54Z)
• MOSAIC: Multi-Subject Personalized Generation via Correspondence-Aware Alignment and Disentanglement [13.100620283631557]
  We present MOSAIC, a representation-centric framework that rethinks multi-subject generation.<n>Our key insight is that multi-subject generation requires precise semantic alignment at the representation level.<n>We propose the semantic correspondence attention loss to enforce precise point-to-point semantic alignment.
  arXiv  Detail & Related papers  (2025-09-02T05:40:07Z)
• Be Yourself: Bounded Attention for Multi-Subject Text-to-Image Generation [60.943159830780154]
  We introduce Bounded Attention, a training-free method for bounding the information flow in the sampling process. We demonstrate that our method empowers the generation of multiple subjects that better align with given prompts and layouts.
  arXiv  Detail & Related papers  (2024-03-25T17:52:07Z)
• RetriBooru: Leakage-Free Retrieval of Conditions from Reference Images for Subject-Driven Generation [30.143033020296183]
  Diffusion-based methods have demonstrated capabilities in generating a diverse array of high-quality images. We propose a multi-level, same-identity dataset RetriBooru, which groups anime characters by both face and cloth identities. We introduce a new concept composition task, where the conditioning encoder learns to retrieve different concepts from several reference images.
  arXiv  Detail & Related papers  (2023-12-05T06:04:16Z)
• SMILE: Semantically-guided Multi-attribute Image and Layout Editing [154.69452301122175]
  Attribute image manipulation has been a very active topic since the introduction of Generative Adversarial Networks (GANs) We present a multimodal representation that handles all attributes, be it guided by random noise or images, while only using the underlying domain information of the target domain. Our method is capable of adding, removing or changing either fine-grained or coarse attributes by using an image as a reference or by exploring the style distribution space.
  arXiv  Detail & Related papers  (2020-10-05T20:15:21Z)

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.