FuseAnyPart: Diffusion-Driven Facial Parts Swapping via Multiple Reference Images
This addresses the need for fine-grained and customized character design in applications like entertainment or digital art, though it is an incremental improvement over existing face swapping techniques.
The paper tackles the problem of swapping individual facial parts in images, which is limited in existing full-face swapping methods, by proposing FuseAnyPart, a diffusion-driven approach that assembles parts from multiple references in latent space and fuses them using a UNet, achieving superior and robust results as validated by experiments.
Facial parts swapping aims to selectively transfer regions of interest from the source image onto the target image while maintaining the rest of the target image unchanged. Most studies on face swapping designed specifically for full-face swapping, are either unable or significantly limited when it comes to swapping individual facial parts, which hinders fine-grained and customized character designs. However, designing such an approach specifically for facial parts swapping is challenged by a reasonable multiple reference feature fusion, which needs to be both efficient and effective. To overcome this challenge, FuseAnyPart is proposed to facilitate the seamless "fuse-any-part" customization of the face. In FuseAnyPart, facial parts from different people are assembled into a complete face in latent space within the Mask-based Fusion Module. Subsequently, the consolidated feature is dispatched to the Addition-based Injection Module for fusion within the UNet of the diffusion model to create novel characters. Extensive experiments qualitatively and quantitatively validate the superiority and robustness of FuseAnyPart. Source codes are available at https://github.com/Thomas-wyh/FuseAnyPart.