2D Instance Editing in 3D Space
This addresses the issue of object consistency and identity preservation in 2D image editing for users of generative models, representing a novel method rather than an incremental improvement.
The paper tackles the problem of inconsistent and identity-losing 2D image editing by introducing a '2D-3D-2D' framework that lifts objects into 3D for editing, then reprojects them, resulting in superior performance over existing methods like DragGAN and DragDiffusion.
Generative models have achieved significant progress in advancing 2D image editing, demonstrating exceptional precision and realism. However, they often struggle with consistency and object identity preservation due to their inherent pixel-manipulation nature. To address this limitation, we introduce a novel "2D-3D-2D" framework. Our approach begins by lifting 2D objects into 3D representation, enabling edits within a physically plausible, rigidity-constrained 3D environment. The edited 3D objects are then reprojected and seamlessly inpainted back into the original 2D image. In contrast to existing 2D editing methods, such as DragGAN and DragDiffusion, our method directly manipulates objects in a 3D environment. Extensive experiments highlight that our framework surpasses previous methods in general performance, delivering highly consistent edits while robustly preserving object identity.