ConsistCompose: Unified Multimodal Layout Control for Image Composition

TL;DR

ConsistCompose introduces a unified multimodal framework embedding layout coordinates into language prompts, enabling precise, layout-controlled multi-instance image generation and outperforming existing methods in spatial accuracy.

Contribution

It presents a novel unified framework and dataset for layout-controlled multimodal image generation, enabling precise spatial control without task-specific modules.

Findings

Improves spatial accuracy over baselines

Preserves identity fidelity in generated images

Demonstrates competitive multimodal understanding

Abstract

Unified multimodal models that couple visual understanding with image generation have advanced rapidly, yet most systems still focus on visual grounding-aligning language with image regions-while their generative counterpart, linguistic-embedded layout-grounded generation (LELG) for layout-controllable multi-instance generation, remains underexplored and limits precise compositional control. We present ConsistCompose, a unified multimodal framework that embeds layout coordinates directly into language prompts, enabling layout-controlled multi-instance image generation from Interleaved Image-Text within a single generative interface. We further construct ConsistCompose3M, a 3.4M multi-instance generation dataset with layout and identity annotations (2.6M text-guided and 0.8M image-guided data pairs) that provides large-scale supervision for layout-conditioned generation. Within this framework, LELG is instantiated through instance-coordinate binding prompts and coordinate-aware classifier-free guidance, which translate linguistic layout cues into precise spatial control without task-specific branches. Experiments on COCO-Position and MS-Bench show that ConsistCompose substantially improves spatial accuracy over layout-controlled baselines while preserving identity fidelity and competitive general multimodal understanding, establishing a unified paradigm for layout-controllable multimodal image generation.