{"slug": "overpainting-localized-context-aware-diffusion-image-editing", "title": "Overpainting: Localized Context-aware Diffusion Image Editing", "summary": "Researchers posted arXiv:2609.10811v1, introducing \"overpainting,\" an image editing operation that controls where an edit occurs via a trimap of white (must edit), gray (may edit), and black (must not edit) pixels. The method adapts a pretrained image editing diffusion model using joint attention and low-rank adaptation across input images with attention-dropout to balance information flow between noise, source, and mask images. The authors also present an automated training data pipeline that generates candidate image pairs with existing language-based editing models, curates them, and extracts a trimap from each usable pair.", "body_md": "arXiv:2609.10811v1 Announce Type: new \nAbstract: We present \"overpainting\", an image editing operation which offers both control over the location of the edit and awareness of the previous content in that location. The overpainted area is given by a trimap, where white-annotated pixels must be edited, gray-annotated pixels may be edited, and black-annotated pixels must not be edited. This enables both precise and loose control, depending on user intent.\n  We implement overpainting by adapting a pretrained image editing diffusion model using a combination of joint attention and low-rank adaption across input images with attention-dropout to balance the information flow between noise, source and mask images. We present a novel, automated, training data generation pipeline that (1) generates a set of candidate image pairs leveraging existing language-based editing models, (2) carefully curates those pairs, and (3) extracts a trimap from each usable pair. We demonstrate the versatility of our overpainting model on a wide range of editing tasks.", "url": "https://wpnews.pro/news/overpainting-localized-context-aware-diffusion-image-editing", "canonical_source": "https://arxiv.org/abs/2609.10811", "published_at": "2026-09-11 04:00:00+00:00", "updated_at": "2026-09-11 04:29:23.664812+00:00", "lang": "en", "topics": ["generative-ai", "computer-vision", "ai-research"], "entities": ["arXiv", "overpainting"], "alternates": {"html": "https://wpnews.pro/news/overpainting-localized-context-aware-diffusion-image-editing", "markdown": "https://wpnews.pro/news/overpainting-localized-context-aware-diffusion-image-editing.md", "text": "https://wpnews.pro/news/overpainting-localized-context-aware-diffusion-image-editing.txt", "jsonld": "https://wpnews.pro/news/overpainting-localized-context-aware-diffusion-image-editing.jsonld"}}