{"slug": "graftsr-grafting-authentic-textures-for-real-world-image-super-resolution-via", "title": "GraftSR: Grafting Authentic Textures for Real-World Image Super-Resolution via Identical-Instance Guidance", "summary": "Researchers propose GraftSR, a diffusion-based image super-resolution framework that uses reference images of the same instance to restore authentic textures, reducing LPIPS by 20.2% over top baselines on the new TexRefSR-Eval benchmark. The framework introduces a dual-mask reference guidance mechanism to handle spatial misalignment, and the team constructed TexRefSR-141K, the first large-scale dataset with reference tuples and complementary spatial masks.", "body_md": "arXiv:2608.25334v1 Announce Type: new\nAbstract: Diffusion-based real-world image super-resolution (SR) achieves impressive perceptual quality but inherently suffers from severe texture hallucination. To overcome this limitation, we propose GraftSR, a texture-reference-guided generative SR framework that leverages reference images of the identical instance to anchor the restoration of authentic textures. However, severe spatial misalignment between low-quality inputs and their references poses significant challenges, often leading to ambiguous transfer targets and background feature leakage. To address these issues, GraftSR employs a novel dual-mask reference guidance mechanism that systematically decouples the cross-view texture injection process. By explicitly isolating what authentic textures to extract from the reference and precisely localizing where to apply them within the target, GraftSR achieves robust texture transfer without relying on brittle spatial alignment. Furthermore, to bridge the critical gap in appropriate training data, we construct TexRefSR-141K, the first large-scale dataset providing high-quality reference tuples equipped with complementary spatial masks. Extensive experiments on our newly established benchmark, TexRefSR-Eval, demonstrate that GraftSR sets a new state-of-the-art. Notably, it reduces LPIPS by 20.2\\% over top-performing baselines, achieving superior reference-faithful restoration.", "url": "https://wpnews.pro/news/graftsr-grafting-authentic-textures-for-real-world-image-super-resolution-via", "canonical_source": "https://arxiv.org/abs/2608.25334", "published_at": "2026-08-27 04:00:00+00:00", "updated_at": "2026-08-27 04:21:56.128319+00:00", "lang": "en", "topics": ["artificial-intelligence", "computer-vision", "generative-ai"], "entities": ["GraftSR", "TexRefSR-141K", "TexRefSR-Eval"], "alternates": {"html": "https://wpnews.pro/news/graftsr-grafting-authentic-textures-for-real-world-image-super-resolution-via", "markdown": "https://wpnews.pro/news/graftsr-grafting-authentic-textures-for-real-world-image-super-resolution-via.md", "text": "https://wpnews.pro/news/graftsr-grafting-authentic-textures-for-real-world-image-super-resolution-via.txt", "jsonld": "https://wpnews.pro/news/graftsr-grafting-authentic-textures-for-real-world-image-super-resolution-via.jsonld"}}