# Hierarchical Concept-to-Appearance Guidance for Multi-Subject Image Generation

arXiv: https://arxiv.org/abs/2602.03448 · HTML: https://mlsanity.com/paper/2602.03448/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models.
- Builds on 25 of its 35 references that are in the map.
- First posted on arXiv 3 Feb 2026.

## Placed in

- Diffusion models / [Image generation, text-to-image and editing](https://mlsanity.com/diffusion/image-generation-text-to-image-and-editing/)

## Builds on (most cited first, 25 of 25)

- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation](https://mlsanity.com/paper/2208.12242/) (2022)
- [InstructPix2Pix: Learning to Follow Image Editing Instructions](https://mlsanity.com/paper/2211.09800/) (2022)
- [An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion](https://mlsanity.com/paper/2208.01618/) (2022)
- [FLUX.1 Kontext: Flow Matching for In-Context Image Generation and Editing in Latent Space](https://mlsanity.com/paper/2506.15742/) (2025)
- [Chameleon: Mixed-Modal Early-Fusion Foundation Models](https://mlsanity.com/paper/2405.09818/) (2024)
- [Emerging Properties in Unified Multimodal Pretraining](https://mlsanity.com/paper/2505.14683/) (2025)
- [Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling](https://mlsanity.com/paper/2501.17811/) (2025)
- [MagicBrush: A Manually Annotated Dataset for Instruction-Guided Image Editing](https://mlsanity.com/paper/2306.10012/) (2023)
- [BLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and Dataset](https://mlsanity.com/paper/2505.09568/) (2025)
- [OminiControl: Minimal and Universal Control for Diffusion Transformer](https://mlsanity.com/paper/2411.15098/) (2024)
- [Emu Edit: Precise Image Editing via Recognition and Generation Tasks](https://mlsanity.com/paper/2311.10089/) (2023)
- [A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence](https://mlsanity.com/paper/2305.15347/) (2023)
- [VIEScore: Towards Explainable Metrics for Conditional Image Synthesis Evaluation](https://mlsanity.com/paper/2312.14867/) (2023)
- [Less-to-More Generalization: Unlocking More Controllability by In-Context Generation](https://mlsanity.com/paper/2504.02160/) (2025)
- [SEED-Data-Edit Technical Report: A Hybrid Dataset for Instructional Image Editing](https://mlsanity.com/paper/2405.04007/) (2024)
- [Telling Left from Right: Identifying Geometry-Aware Semantic Correspondence](https://mlsanity.com/paper/2311.17034/) (2023)
- [DreamO: A Unified Framework for Image Customization](https://mlsanity.com/paper/2504.16915/) (2025)
- [SFNet: Learning Object-aware Semantic Correspondence](https://mlsanity.com/paper/1904.01810/) (2019)
- [XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT Modulation](https://mlsanity.com/paper/2506.21416/) (2025)
- [Reference-Based Sketch Image Colorization using Augmented-Self Reference and Dense Semantic Correspondence](https://mlsanity.com/paper/2005.05207/) (2020)
- [OneCAT: Decoder-Only Auto-Regressive Model for Unified Understanding and Generation](https://mlsanity.com/paper/2509.03498/) (2025)
- [RegionDrag: Fast Region-Based Image Editing with Diffusion Models](https://mlsanity.com/paper/2407.18247/) (2024)
- [UMO: Scaling Multi-Identity Consistency for Image Customization via Matching Reward](https://mlsanity.com/paper/2509.06818/) (2025)
- [MOSAIC: Multi-Subject Personalized Generation via Correspondence-Aware Alignment and Disentanglement](https://mlsanity.com/paper/2509.01977/) (2025)

## Built on by (most cited first, 1 of 1)

- [Squeezing Capacity from Multimodal Large Language Models for Subject-driven Generation](https://mlsanity.com/paper/2605.26111/) (2026)

## Abstract (arXiv)

Multi-subject image generation aims to synthesize images that faithfully preserve the identities of multiple reference subjects while following textual instructions. However, existing methods often suffer from identity inconsistency and limited compositional control, as they rely on diffusion models to implicitly associate text prompts with reference images. In this work, we propose Hierarchical Concept-to-Appearance Guidance (CAG), a framework that provides explicit, structured supervision from high-level concepts to fine-grained appearances. At the conceptual level, we introduce a VAE dropout training strategy that randomly omits reference VAE features, encouraging the model to rely more on robust semantic signals from a Visual Language Model (VLM) and thereby promoting consistent concept-level generation in the absence of complete appearance cues. At the appearance level, we integrate the VLM-derived correspondences into a correspondence-aware masked attention module within the Diffusion Transformer (DiT). This module restricts each text token to attend only to its matched reference regions, ensuring precise attribute binding and reliable multi-subject composition. Extensive experiments demonstrate that our method achieves state-of-the-art performance on the multi-subject image generation, substantially improving prompt following and subject consistency.