We introduce <strong>MirrorPPR</strong>, a novel framework for <strong>Exemplar-Based Portrait Photo Retouching</strong>, which captures subtle retouching operations from an exemplar pair and transfers them to a new query image. To support this task, we propose a <strong>data self-augmentation</strong> paradigm and construct MirrorPPR47M, a large-scale dataset with over 47 million retouched pairs. Extensive experiments demonstrate that MirrorPPR significantly outperforms existing baselines in both retouching quality and identity preservation.</p>\n","updatedAt":"2026-06-30T22:27:41.209Z","author":{"_id":"6787b44e9e6186bec4033b21","avatarUrl":"/avatars/db0caeffb0a62dee79df2e9b7f1a2121.svg","fullname":"lucky-lzh","name":"lucky-lzh","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8336028456687927},"editors":["lucky-lzh"],"editorAvatarUrls":["/avatars/db0caeffb0a62dee79df2e9b7f1a2121.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.29308","authors":[{"_id":"6a436465763f63ca3757ead6","name":"Zhihong Liu","hidden":false},{"_id":"6a436465763f63ca3757ead7","name":"Zheng Li","hidden":false},{"_id":"6a436465763f63ca3757ead8","name":"Jiachun Jin","hidden":false},{"_id":"6a436465763f63ca3757ead9","name":"Siqi Kou","hidden":false},{"_id":"6a436465763f63ca3757eada","name":"Yitao Jian","hidden":false},{"_id":"6a436465763f63ca3757eadb","name":"Fengpei Yu","hidden":false},{"_id":"6a436465763f63ca3757eadc","name":"Zhijie Deng","hidden":false}],"publishedAt":"2026-06-28T00:00:00.000Z","submittedOnDailyAt":"2026-06-30T00:00:00.000Z","title":"MirrorPPR: Exemplar-Based Portrait Photo Retouching","submittedOnDailyBy":{"_id":"6787b44e9e6186bec4033b21","avatarUrl":"/avatars/db0caeffb0a62dee79df2e9b7f1a2121.svg","isPro":false,"fullname":"lucky-lzh","user":"lucky-lzh","type":"user","name":"lucky-lzh"},"summary":"While text-guided image editing has made remarkable progress, it remains limited in structural portrait retouching. Textual descriptions struggle to convey fine-grained changes to facial features and body proportions. To address this gap, we introduce Exemplar-Based Portrait Photo Retouching, where the model is given an exemplar pair and tasked with inferring and applying the same retouching operations to a new query image. Existing exemplar-based editing methods primarily focus on tasks with pronounced visual transformations. In contrast, structural portrait retouching involves extremely delicate and localized modifications, making accurate extraction and transfer of these edits challenging. To tackle this, we propose MirrorPPR, a novel framework designed to capture and transfer subtle structural retouching operations. Our method uses a Retouching Operation Extractor to capture the subtle differences from the exemplar pair. The extracted representations are then injected into a pre-trained Diffusion Transformer (DiT) through a connector and Low-Rank Adaptation (LoRA) modules. Furthermore, constructing perfectly aligned cross-identity training pairs is severely hindered by operation misalignment. To overcome this, we propose an advanced data self-augmentation paradigm that ensures strictly aligned retouching operations. To alleviate data scarcity and support this novel task, we introduce MirrorPPR47M, a large-scale dataset with over 47 million retouched pairs. By structuring the dataset into simulated and professional subsets, we enable progressive curriculum learning to smoothly optimize the network. Extensive experiments demonstrate that MirrorPPR significantly outperforms existing baselines in both retouching quality and identity preservation. The project page is available at https://sjtu-deng-lab.github.io/MirrorPPR.","upvotes":1,"discussionId":"6a436465763f63ca3757eadd","projectPage":"https://sjtu-deng-lab.github.io/MirrorPPR","githubRepo":"https://github.com/SJTU-DENG-Lab/MirrorPPR","githubRepoAddedBy":"user","ai_summary":"Exemplar-based portrait retouching framework using Diffusion Transformer with LoRA adaptation and self-augmented training data achieves superior quality and identity preservation.","ai_keywords":["exemplar-based editing","Diffusion Transformer","DiT","Low-Rank Adaptation","LoRA","retouching operation extractor","cross-identity training pairs","data self-augmentation","curriculum learning","MirrorPPR47M dataset"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":3,"organization":{"_id":"673d5fe8d031224e947dc235","name":"SJTU-DENG-Lab","fullname":"DENG Lab @ SJTU","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/64bba541da140e461924dfed/_WPqM9jCqIIkS73aTeZP-.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6787b44e9e6186bec4033b21","avatarUrl":"/avatars/db0caeffb0a62dee79df2e9b7f1a2121.svg","isPro":false,"fullname":"lucky-lzh","user":"lucky-lzh","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"673d5fe8d031224e947dc235","name":"SJTU-DENG-Lab","fullname":"DENG Lab @ SJTU","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/64bba541da140e461924dfed/_WPqM9jCqIIkS73aTeZP-.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2606/2606.29308.md","query":{}}">
MirrorPPR: Exemplar-Based Portrait Photo Retouching
Abstract
Exemplar-based portrait retouching framework using Diffusion Transformer with LoRA adaptation and self-augmented training data achieves superior quality and identity preservation.
While text-guided image editing has made remarkable progress, it remains limited in structural portrait retouching. Textual descriptions struggle to convey fine-grained changes to facial features and body proportions. To address this gap, we introduce Exemplar-Based Portrait Photo Retouching, where the model is given an exemplar pair and tasked with inferring and applying the same retouching operations to a new query image. Existing exemplar-based editing methods primarily focus on tasks with pronounced visual transformations. In contrast, structural portrait retouching involves extremely delicate and localized modifications, making accurate extraction and transfer of these edits challenging. To tackle this, we propose MirrorPPR, a novel framework designed to capture and transfer subtle structural retouching operations. Our method uses a Retouching Operation Extractor to capture the subtle differences from the exemplar pair. The extracted representations are then injected into a pre-trained Diffusion Transformer (DiT) through a connector and Low-Rank Adaptation (LoRA) modules. Furthermore, constructing perfectly aligned cross-identity training pairs is severely hindered by operation misalignment. To overcome this, we propose an advanced data self-augmentation paradigm that ensures strictly aligned retouching operations. To alleviate data scarcity and support this novel task, we introduce MirrorPPR47M, a large-scale dataset with over 47 million retouched pairs. By structuring the dataset into simulated and professional subsets, we enable progressive curriculum learning to smoothly optimize the network. Extensive experiments demonstrate that MirrorPPR significantly outperforms existing baselines in both retouching quality and identity preservation. The project page is available at https://sjtu-deng-lab.github.io/MirrorPPR.
Community
We introduce MirrorPPR, a novel framework for Exemplar-Based Portrait Photo Retouching, which captures subtle retouching operations from an exemplar pair and transfers them to a new query image. To support this task, we propose a data self-augmentation paradigm and construct MirrorPPR47M, a large-scale dataset with over 47 million retouched pairs. Extensive experiments demonstrate that MirrorPPR significantly outperforms existing baselines in both retouching quality and identity preservation.
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/2606.29308 in a model README.md to link it from this page.
Cite arxiv.org/abs/2606.29308 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2606.29308 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.