MOSAIC: Multi-Subject Personalized Generation via Correspondence-Aware Alignment and Disentanglement

2025-09-02

AI Summary

Multi-subject personalized image generation suffers from identity confusion and semantic misalignment, primarily due to the absence of explicit modeling of inter-subject interactions within shared representation spaces. To address this, we propose a synergistic framework integrating semantic alignment and orthogonal feature disentanglement. We first introduce SemAlign-MS—the first fine-grained multi-subject semantic correspondence dataset. Second, we design a semantic correspondence attention loss to enable precise cross-subject region-level alignment. Third, we propose a multi-reference disentanglement loss coupled with orthogonal subspace constraints to isolate subject-specific identity features. Our method is integrated into a diffusion-based generative architecture. Experiments demonstrate state-of-the-art performance across multiple benchmarks, enabling high-fidelity synthesis of ≥4 subjects simultaneously. Our approach significantly improves identity fidelity (+12.7% ID-Retrieval) and semantic consistency (+9.3% CLIP-Score).

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Multi-subject personalized generation presents unique challenges in maintaining identity fidelity and semantic coherence when synthesizing images conditioned on multiple reference subjects. Existing methods often suffer from identity blending and attribute leakage due to inadequate modeling of how different subjects should interact within shared representation spaces. We present MOSAIC, a representation-centric framework that rethinks multi-subject generation through explicit semantic correspondence and orthogonal feature disentanglement. Our key insight is that multi-subject generation requires precise semantic alignment at the representation level - knowing exactly which regions in the generated image should attend to which parts of each reference. To enable this, we introduce SemAlign-MS, a meticulously annotated dataset providing fine-grained semantic correspondences between multiple reference subjects and target images, previously unavailable in this domain. Building on this foundation, we propose the semantic correspondence attention loss to enforce precise point-to-point semantic alignment, ensuring high consistency from each reference to its designated regions. Furthermore, we develop the multi-reference disentanglement loss to push different subjects into orthogonal attention subspaces, preventing feature interference while preserving individual identity characteristics. Extensive experiments demonstrate that MOSAIC achieves state-of-the-art performance on multiple benchmarks. Notably, while existing methods typically degrade beyond 3 subjects, MOSAIC maintains high fidelity with 4+ reference subjects, opening new possibilities for complex multi-subject synthesis applications.

Problem

• Maintaining identity fidelity in multi-subject image generation
• Preventing identity blending and attribute leakage issues
• Enabling precise semantic alignment for multiple reference subjects

Innovation

• Semantic correspondence attention loss for precise alignment
• Multi-reference disentanglement loss for orthogonal subspaces
• SemAlign-MS dataset enables fine-grained semantic correspondences