PreciseControl: Enhancing Text-To-Image Diffusion Models with Fine-Grained Attribute Control

TL;DR

PreciseControl introduces a method that combines StyleGAN's rich face prior with text-to-image diffusion models, enabling precise, fine-grained facial attribute editing while maintaining identity and coarse control.

Contribution

The paper proposes a novel approach using StyleGAN's $ ext{W+}$ space to condition T2I models, allowing detailed facial attribute manipulation with improved inversion and editing capabilities.

Findings

Enhanced face inversion with identity preservation

Enables smooth, fine-grained attribute editing

Supports multi-person image composition

Abstract

Recently, we have seen a surge of personalization methods for text-to-image (T2I) diffusion models to learn a concept using a few images. Existing approaches, when used for face personalization, suffer to achieve convincing inversion with identity preservation and rely on semantic text-based editing of the generated face. However, a more fine-grained control is desired for facial attribute editing, which is challenging to achieve solely with text prompts. In contrast, StyleGAN models learn a rich face prior and enable smooth control towards fine-grained attribute editing by latent manipulation. This work uses the disentangled $\mathcal{W+}$ space of StyleGANs to condition the T2I model. This approach allows us to precisely manipulate facial attributes, such as smoothly introducing a smile, while preserving the existing coarse text-based control inherent in T2I models. To enable conditioning of the T2I model on the $\mathcal{W+}$ space, we train a latent mapper to translate latent codes from $\mathcal{W+}$ to the token embedding space of the T2I model. The proposed approach excels in the precise inversion of face images with attribute preservation and facilitates continuous control for fine-grained attribute editing. Furthermore, our approach can be readily extended to generate compositions involving multiple individuals. We perform extensive experiments to validate our method for face personalization and fine-grained attribute editing.