LayerComposer: Multi-Human Personalized Generation via Layered Canvas

TL;DR

LayerComposer is a scalable, interactive framework for multi-human personalized image generation that offers precise spatial control, occlusion-free composition, and identity preservation through layered representation and novel training strategies.

Contribution

It introduces a layered canvas representation, transparent latent pruning, and layerwise cross-reference training to enhance multi-human personalized image generation.

Findings

Achieves superior spatial control and composition coherence.

Maintains identity preservation in multi-human scenarios.

Demonstrates scalability with multiple subjects.

Abstract

Despite their impressive visual fidelity, existing personalized image generators lack interactive control over spatial composition and scale poorly to multiple humans. To address these limitations, we present LayerComposer, an interactive and scalable framework for multi-human personalized generation. Inspired by professional image-editing software, LayerComposer provides intuitive reference-based human injection, allowing users to place and resize multiple subjects directly on a layered digital canvas to guide personalized generation. The core of our approach is the layered canvas, a novel representation where each subject is placed on a distinct layer, enabling interactive and occlusion-free composition. We further introduce a transparent latent pruning mechanism that improves scalability by decoupling computational cost from the number of subjects, and a layerwise cross-reference training strategy that mitigates copy-paste artifacts. Extensive experiments demonstrate that LayerComposer achieves superior spatial control, coherent composition, and identity preservation compared to state-of-the-art methods in multi-human personalized image generation.