Perceptual Similarity guidance and text guidance optimization for Editing Real Images using Guided Diffusion Models (2312.06680v1)
Abstract: When using a diffusion model for image editing, there are times when the modified image can differ greatly from the source. To address this, we apply a dual-guidance approach to maintain high fidelity to the original in areas that are not altered. First, we employ text-guided optimization, using text embeddings to direct latent space and classifier-free guidance. Second, we use perceptual similarity guidance, optimizing latent vectors with posterior sampling via Tweedie formula during the reverse process. This method ensures the realistic rendering of both the edited elements and the preservation of the unedited parts of the original image.
- Ilvr: Conditioning method for denoising diffusion probabilistic models, 2021.
- Diffedit: Diffusion-based semantic image editing with mask guidance, 2022.
- High-fidelity and arbitrary face editing, 2021.
- Prompt-to-prompt image editing with cross attention control, 2022.
- Clipscore: A reference-free evaluation metric for image captioning, 2022.
- Classifier-free diffusion guidance, 2022.
- Enhancing diffusion-based image synthesis with robust classifier guidance, 2023.
- Details or artifacts: A locally discriminative learning approach to realistic image super-resolution, 2022.
- Repaint: Inpainting using denoising diffusion probabilistic models, 2022.
- Sdedit: Guided image synthesis and editing with stochastic differential equations, 2022.
- Null-text inversion for editing real images using guided diffusion models, 2022.
- Lanit: Language-driven image-to-image translation for unlabeled data, 2023.
- High-resolution image synthesis with latent diffusion models, 2022.
- Knn-diffusion: Image generation via large-scale retrieval, 2022.
- Adaint: Learning adaptive intervals for 3d lookup tables on real-time image enhancement, 2022.
- The unreasonable effectiveness of deep features as a perceptual metric, 2018.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.
