TL;DR: A VAE-free pixel-space DiT for unified image restoration. It applies flow matching to patchified RGB, uses adaptive DINO guidance, and is finetuned for one-step inference with ~50M parameters.</p>\n","updatedAt":"2026-08-19T08:04:23.134Z","author":{"_id":"5f1158120c833276f61f1a84","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1608042047613-5f1158120c833276f61f1a84.jpeg","fullname":"Niels Rogge","name":"nielsr","type":"user","isPro":false,"isHf":true,"isHfAdmin":false,"isMod":false,"followerCount":1278,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7754538655281067},"editors":["nielsr"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1608042047613-5f1158120c833276f61f1a84.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.16793","authors":[{"_id":"6a85637a536bdd3bdd48f96b","name":"Lingchen Sun","hidden":false},{"_id":"6a85637a536bdd3bdd48f96c","name":"Rongyuan Wu","hidden":false},{"_id":"6a85637a536bdd3bdd48f96d","name":"Xiangtao Kong","hidden":false},{"_id":"6a85637a536bdd3bdd48f96e","name":"Jixin Zhao","hidden":false},{"_id":"6a85637a536bdd3bdd48f96f","name":"Qiaosi Yi","hidden":false},{"_id":"6a85637a536bdd3bdd48f970","name":"Yujing Sun","hidden":false},{"_id":"6a85637a536bdd3bdd48f971","name":"Shuaizheng Liu","hidden":false},{"_id":"6a85637a536bdd3bdd48f972","name":"Zhengqiang Zhang","hidden":false},{"_id":"6a85637a536bdd3bdd48f973","name":"Lei Zhang","hidden":false}],"publishedAt":"2026-08-17T00:00:00.000Z","submittedOnDailyAt":"2026-08-19T00:00:00.000Z","title":"PixRestore: Unified Image Restoration via Pixel Diffusion Transformer","submittedOnDailyBy":{"_id":"5f1158120c833276f61f1a84","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1608042047613-5f1158120c833276f61f1a84.jpeg","isPro":false,"fullname":"Niels Rogge","user":"nielsr","type":"user","name":"nielsr"},"summary":"Unified image restoration (UIR) aims to recover high-quality (HQ) content from low-quality (LQ) images with different degradations using a single model. Most recent methods adapt large pretrained text-to-image (T2I) latent diffusion models for their strong capacity and generative priors. However, the variational autoencoder (VAE) in latent T2I models may discard restoration-sensitive details, while the open-ended synthesis prior can introduce content-inconsistent artifacts. We present PixRestore, a VAE-free pixel-space Diffusion Transformer (DiT) for UIR, where the diffusion backbone is trained entirely from scratch, without relying on T2I pretraining. PixRestore performs flow matching directly on patchified pixels, preserving fine-grained details while keeping the token sequence tractable. To adapt to different degradations, PixRestore learns to predict the reliability of layer features using LQ--HQ DINO feature similarity. Features from more reliable layers are fused as dense conditioning, while less reliable layers receive stronger HQ-feature supervision to encourage degradation removal. We train PixRestore on a large-scale corpus of diverse scenes and degradations, and further finetune it into a one-step generator using DINO-based adversarial objectives for efficient inference. Experiments on public benchmarks and real-world test sets show that, with only about 50M parameters and single-step inference, PixRestore achieves the best overall fidelity, perceptual quality, and robustness to degradations among competing UIR models while being far more efficient. Larger PixRestore variants can further boost performance, demonstrating the scalability of our pixel-space design. Code and the curated benchmark can be found at https://github.com/csslc/PixRestore.","upvotes":0,"discussionId":"6a85637a536bdd3bdd48f974","githubRepo":"https://github.com/csslc/PixRestore","githubRepoAddedBy":"admin","ai_summary":"PixRestore is a compact, VAE-free pixel-space diffusion transformer trained from scratch for unified image restoration, using flow matching on patchified pixels, DINO-based reliability-guided feature fusion, and adversarial fine-tuning to a one-step generator for efficient high-fidelity inference.","ai_keywords":["unified image restoration","text-to-image latent diffusion models","variational autoencoder","Diffusion Transformer","flow matching","patchified pixels","DINO feature similarity","dense conditioning","adversarial objectives","one-step generator"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":26},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[],"acceptLanguages":["en"],"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.16793.md","query":{}}">
PixRestore: Unified Image Restoration via Pixel Diffusion Transformer
Abstract
PixRestore is a compact, VAE-free pixel-space diffusion transformer trained from scratch for unified image restoration, using flow matching on patchified pixels, DINO-based reliability-guided feature fusion, and adversarial fine-tuning to a one-step generator for efficient high-fidelity inference.
Unified image restoration (UIR) aims to recover high-quality (HQ) content from low-quality (LQ) images with different degradations using a single model. Most recent methods adapt large pretrained text-to-image (T2I) latent diffusion models for their strong capacity and generative priors. However, the variational autoencoder (VAE) in latent T2I models may discard restoration-sensitive details, while the open-ended synthesis prior can introduce content-inconsistent artifacts. We present PixRestore, a VAE-free pixel-space Diffusion Transformer (DiT) for UIR, where the diffusion backbone is trained entirely from scratch, without relying on T2I pretraining. PixRestore performs flow matching directly on patchified pixels, preserving fine-grained details while keeping the token sequence tractable. To adapt to different degradations, PixRestore learns to predict the reliability of layer features using LQ--HQ DINO feature similarity. Features from more reliable layers are fused as dense conditioning, while less reliable layers receive stronger HQ-feature supervision to encourage degradation removal. We train PixRestore on a large-scale corpus of diverse scenes and degradations, and further finetune it into a one-step generator using DINO-based adversarial objectives for efficient inference. Experiments on public benchmarks and real-world test sets show that, with only about 50M parameters and single-step inference, PixRestore achieves the best overall fidelity, perceptual quality, and robustness to degradations among competing UIR models while being far more efficient. Larger PixRestore variants can further boost performance, demonstrating the scalability of our pixel-space design. Code and the curated benchmark can be found at https://github.com/csslc/PixRestore.
Community
TL;DR: A VAE-free pixel-space DiT for unified image restoration. It applies flow matching to patchified RGB, uses adaptive DINO guidance, and is finetuned for one-step inference with ~50M parameters.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2608.16793 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.16793 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.16793 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.