Paper ID: 2411.18068

PersonaCraft: Personalized Full-Body Image Synthesis for Multiple Identities from Single References Using 3D-Model-Conditioned Diffusion

Gwanghyun Kim, Suh Yoon Jeon, Seunggyu Lee, Se Young Chun

Personalized image generation has been significantly advanced, enabling the creation of highly realistic and customized images. However, existing methods often struggle with generating images of multiple people due to occlusions and fail to accurately personalize full-body shapes. In this paper, we propose PersonaCraft, a novel approach that combines diffusion models with 3D human modeling to address these limitations. Our method effectively manages occlusions by incorporating 3D-aware pose conditioning with SMPLx-ControlNet and accurately personalizes human full-body shapes through SMPLx fitting. Additionally, PersonaCraft enables user-defined body shape adjustments, adding flexibility for individual body customization. Experimental results demonstrate the superior performance of PersonaCraft in generating high-quality, realistic images of multiple individuals while resolving occlusion issues, thus establishing a new standard for multi-person personalized image synthesis. Project page: this https URL

Submitted: Nov 27, 2024