CVDec 13, 2023

Compositional Inversion for Stable Diffusion Models

arXiv:2312.08048v326 citationsh-index: 7Has CodeAAAI
AI Analysis

This addresses a specific issue in personalized image generation for users of diffusion models, offering an incremental improvement to existing inversion techniques.

The paper tackles the overfitting problem in inversion methods for Stable Diffusion models, where inverted concepts dominate and suppress others, by proposing a method that guides inversion towards the core distribution and uses spatial regularization to balance attention, resulting in more diverse and balanced concept compositions in synthesized images.

Inversion methods, such as Textual Inversion, generate personalized images by incorporating concepts of interest provided by user images. However, existing methods often suffer from overfitting issues, where the dominant presence of inverted concepts leads to the absence of other desired concepts. It stems from the fact that during inversion, the irrelevant semantics in the user images are also encoded, forcing the inverted concepts to occupy locations far from the core distribution in the embedding space. To address this issue, we propose a method that guides the inversion process towards the core distribution for compositional embeddings. Additionally, we introduce a spatial regularization approach to balance the attention on the concepts being composed. Our method is designed as a post-training approach and can be seamlessly integrated with other inversion methods. Experimental results demonstrate the effectiveness of our proposed approach in mitigating the overfitting problem and generating more diverse and balanced compositions of concepts in the synthesized images. The source code is available at https://github.com/zhangxulu1996/Compositional-Inversion.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes