ComposeAnyone: Controllable Layout-to-Human Generation with Decoupled Multimodal Conditions

ComposeAnyone is capable of generating high-quality human images with decoupled multimodal conditions, such as captions, reference images, and hand-drawn layouts.

---