Stylistic Attribute Control in Latent Diffusion Models
For users of text-to-image diffusion models, this work addresses the challenge of fine-grained stylistic control without unintended content changes, offering a more precise editing tool.
The paper tackles precise control over stylistic attributes in text-to-image diffusion models, proposing a method that learns disentangled editing directions from synthetic datasets and uses guidance composition to preserve semantics. The approach achieves more precise and continuously adjustable stylistic modifications compared to current text-based editing techniques.
Text-to-image diffusion models have revolutionized image synthesis and editing, but precise control over stylistic attributes remains a challenge, often causing unintended content modifications. We propose an approach for fine-grained parametric control of stylistic attributes in latent diffusion models by learning disentangled editing directions from synthetic datasets. We use guidance composition to close the domain gap between stylistically finetuned and foundation models, preserving the original image semantics while applying stylistic adjustments. To ensure consistent edits, we introduce a training regularization loss and enhance DDIM inversion with optimized null-conditional embeddings for real image editing. We validate our approach by learning from stylistically filtered synthetic datasets varying a range of stylistic attributes, including outlines, local contrast, watercolorization effects, and geometric patterns. Our evaluations demonstrate that compared to current text-based editing techniques, our method offers well-integrated, more precise and continuously adjustable stylistic modifications.