Interesting work on <strong>video reflection removal</strong>, a problem that has received much less attention than single-image reflection removal.</p>\n<p>The paper proposes a closed-loop approach combining <strong>physics-grounded reflection synthesis, diffusion-based video dereflection, and dedicated benchmark design</strong>. In particular, S2R-Synthesis generates realistic reflected videos for training, S2R-Removal adapts a pretrained video diffusion prior for one-step reflection removal, and S2R-Bench provides a dedicated benchmark for evaluating video dereflection.</p>\n<p>The combination of <strong>realistic data synthesis + video diffusion + benchmark design</strong> makes this a promising step toward more practical video reflection removal.</p>\n<p><strong>Code and pretrained weights will be released soon.</strong> Stay tuned!</p>\n","updatedAt":"2026-08-13T02:24:03.324Z","author":{"_id":"649176436cadae13f22d014b","avatarUrl":"/avatars/fa94a8806499a5bd532ca7030e904cd1.svg","fullname":"Higher","name":"HigherHu","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8182305097579956},"editors":["HigherHu"],"editorAvatarUrls":["/avatars/fa94a8806499a5bd532ca7030e904cd1.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.11562","authors":[{"_id":"6a7d27470ac8bee77474ee64","name":"Zepeng Wang","hidden":false},{"_id":"6a7d27470ac8bee77474ee65","name":"Jiagao Hu","hidden":false},{"_id":"6a7d27470ac8bee77474ee66","name":"Fuhao Li","hidden":false},{"_id":"6a7d27470ac8bee77474ee67","name":"Yuxuan Chen","hidden":false},{"_id":"6a7d27470ac8bee77474ee68","name":"Fei Wang","hidden":false},{"_id":"6a7d27470ac8bee77474ee69","name":"Daiguo Zhou","hidden":false}],"publishedAt":"2026-08-12T00:00:00.000Z","submittedOnDailyAt":"2026-08-13T00:00:00.000Z","title":"From Synthesis to Removal: Physics-Grounded Reflection Simulation and Diffusion-Based Video Dereflection","submittedOnDailyBy":{"_id":"649176436cadae13f22d014b","avatarUrl":"/avatars/fa94a8806499a5bd532ca7030e904cd1.svg","isPro":false,"fullname":"Higher","user":"HigherHu","type":"user","name":"HigherHu"},"summary":"Videos captured through glass often contain reflections that degrade visual quality and interfere with downstream vision tasks. Although single-image reflection removal has been extensively studied, video reflection removal remains largely underexplored due to the lack of paired video data, temporally coherent removal models, and dedicated evaluation benchmarks. We present a closed-loop framework that unifies physics-grounded reflection simulation, diffusion-based video dereflection, and benchmark evaluation. Our S2R-Synthesis pipeline generates paired reflected and reflection-free videos by performing physics-grounded augmentation in the structure space and rendering realistic reflected videos with a trained video diffusion renderer; the augmentation models key glass-related effects including roughness-induced blur, thickness-induced ghosting, and reflectance variation. Based on the synthesized data, we introduce S2R-Removal, the first diffusion-based video reflection removal model, which adapts a pretrained video diffusion prior through reflection-aware latent adaptation and one-step pixel-geometric refinement, recovering the clean transmission in a single denoising step. We further build S2R-Bench, the first benchmark for video reflection removal, supporting both full-reference evaluation and real-world human perceptual assessment. Experiments on S2R-Bench and multiple public image benchmarks demonstrate state-of-the-art performance and faster inference than even non-diffusion baselines, and validate the effectiveness of S2R-Synthesis. Project page: https://codingwzp.github.io/VideoDereflection_S2R.","upvotes":6,"discussionId":"6a7d27470ac8bee77474ee6a","projectPage":"https://codingwzp.github.io/VideoDereflection_S2R/","ai_summary":"A closed-loop framework combining physics-based video synthesis, diffusion-based video dereflection, and a new benchmark achieves state-of-the-art video reflection removal with fast inference.","ai_keywords":["diffusion-based video dereflection","physics-grounded reflection simulation","video diffusion renderer","reflection-aware latent adaptation","one-step pixel-geometric refinement","S2R-Synthesis","S2R-Removal","S2R-Bench"],"ai_summary_model":"thinkingmachines/Inkling-Small","organization":{"_id":"6821ba7e5a7efab94a235406","name":"xiaomi-research","fullname":"Xiaomi Research","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/673735e4373ad40af7f81ea1/DR4m0bz2Du1l0Z8Txg351.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"66447562e7ca43b97c714cb1","avatarUrl":"/avatars/dc0869849b52650fe4ac7a2c6121a91a.svg","isPro":false,"fullname":"wangzepeng","user":"codingwzp","type":"user"},{"_id":"66a849128d74567a61afa114","avatarUrl":"/avatars/e8bf027eed0744e11a48152c9275326d.svg","isPro":false,"fullname":"Yuxuan Chen","user":"Saandsa","type":"user"},{"_id":"65c4eb7cd1dcbd30d86febec","avatarUrl":"/avatars/001c8f02e8ce794b2c21883628b2da72.svg","isPro":false,"fullname":"free-bit","user":"free-bit","type":"user"},{"_id":"6603f231b4344a2b07d5bc05","avatarUrl":"/avatars/da5d570d1e5083a711cdfeca0c6afbad.svg","isPro":false,"fullname":"Ruoyu Wang","user":"dwawayu","type":"user"},{"_id":"64e9f5bbf494f8b2a0670eee","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/rM0q9i3DXc7nVY6MDjiCY.jpeg","isPro":false,"fullname":"Rijusmit Biswas","user":"Phantomcloak19","type":"user"},{"_id":"649176436cadae13f22d014b","avatarUrl":"/avatars/fa94a8806499a5bd532ca7030e904cd1.svg","isPro":false,"fullname":"Higher","user":"HigherHu","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6821ba7e5a7efab94a235406","name":"xiaomi-research","fullname":"Xiaomi Research","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/673735e4373ad40af7f81ea1/DR4m0bz2Du1l0Z8Txg351.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.11562.md","query":{}}">
From Synthesis to Removal: Physics-Grounded Reflection Simulation and Diffusion-Based Video Dereflection
Published on Aug 12
· Submitted by Higher on Aug 13 Abstract
A closed-loop framework combining physics-based video synthesis, diffusion-based video dereflection, and a new benchmark achieves state-of-the-art video reflection removal with fast inference.
Videos captured through glass often contain reflections that degrade visual quality and interfere with downstream vision tasks. Although single-image reflection removal has been extensively studied, video reflection removal remains largely underexplored due to the lack of paired video data, temporally coherent removal models, and dedicated evaluation benchmarks. We present a closed-loop framework that unifies physics-grounded reflection simulation, diffusion-based video dereflection, and benchmark evaluation. Our S2R-Synthesis pipeline generates paired reflected and reflection-free videos by performing physics-grounded augmentation in the structure space and rendering realistic reflected videos with a trained video diffusion renderer; the augmentation models key glass-related effects including roughness-induced blur, thickness-induced ghosting, and reflectance variation. Based on the synthesized data, we introduce S2R-Removal, the first diffusion-based video reflection removal model, which adapts a pretrained video diffusion prior through reflection-aware latent adaptation and one-step pixel-geometric refinement, recovering the clean transmission in a single denoising step. We further build S2R-Bench, the first benchmark for video reflection removal, supporting both full-reference evaluation and real-world human perceptual assessment. Experiments on S2R-Bench and multiple public image benchmarks demonstrate state-of-the-art performance and faster inference than even non-diffusion baselines, and validate the effectiveness of S2R-Synthesis. Project page: https://codingwzp.github.io/VideoDereflection_S2R.
Community
Interesting work on video reflection removal, a problem that has received much less attention than single-image reflection removal.
The paper proposes a closed-loop approach combining physics-grounded reflection synthesis, diffusion-based video dereflection, and dedicated benchmark design. In particular, S2R-Synthesis generates realistic reflected videos for training, S2R-Removal adapts a pretrained video diffusion prior for one-step reflection removal, and S2R-Bench provides a dedicated benchmark for evaluating video dereflection.
The combination of realistic data synthesis + video diffusion + benchmark design makes this a promising step toward more practical video reflection removal.
Code and pretrained weights will be released soon. Stay tuned!
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/2608.11562 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.11562 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.11562 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.