Identity Encoder for Personalized Diffusion
This reduces overhead for personalized image generation in applications like image enhancement and video conferences, though it is incremental over existing encoder-based approaches.
The paper tackles the computational and data inefficiency of fine-tuning separate models for personalized image generation by proposing an identity encoder that extracts representations from a few reference images, enabling generation for unseen identities. Empirical results show it outperforms fine-tuning baselines in generation and reconstruction, with user preference over 95% of the time.
Many applications can benefit from personalized image generation models, including image enhancement, video conferences, just to name a few. Existing works achieved personalization by fine-tuning one model for each person. While being successful, this approach incurs additional computation and storage overhead for each new identity. Furthermore, it usually expects tens or hundreds of examples per identity to achieve the best performance. To overcome these challenges, we propose an encoder-based approach for personalization. We learn an identity encoder which can extract an identity representation from a set of reference images of a subject, together with a diffusion generator that can generate new images of the subject conditioned on the identity representation. Once being trained, the model can be used to generate images of arbitrary identities given a few examples even if the model hasn't been trained on the identity. Our approach greatly reduces the overhead for personalized image generation and is more applicable in many potential applications. Empirical results show that our approach consistently outperforms existing fine-tuning based approach in both image generation and reconstruction, and the outputs is preferred by users more than 95% of the time compared with the best performing baseline.