KV-Edit: Training-Free Image Editing for Precise Background Preservation
This addresses the problem of precise background preservation in image editing for users, offering a novel, efficient solution without the need for training.
The paper tackles the challenge of background consistency in image editing by proposing KV-Edit, a training-free method that uses KV cache in DiTs to preserve background tokens, achieving significant improvements in background and image quality over existing approaches, including training-based methods.
Background consistency remains a significant challenge in image editing tasks. Despite extensive developments, existing works still face a trade-off between maintaining similarity to the original image and generating content that aligns with the target. Here, we propose KV-Edit, a training-free approach that uses KV cache in DiTs to maintain background consistency, where background tokens are preserved rather than regenerated, eliminating the need for complex mechanisms or expensive training, ultimately generating new content that seamlessly integrates with the background within user-provided regions. We further explore the memory consumption of the KV cache during editing and optimize the space complexity to $O(1)$ using an inversion-free method. Our approach is compatible with any DiT-based generative model without additional training. Experiments demonstrate that KV-Edit significantly outperforms existing approaches in terms of both background and image quality, even surpassing training-based methods. Project webpage is available at https://xilluill.github.io/projectpages/KV-Edit