Project page: <a href=\"https://cvlab-kaist.github.io/GARD/\" rel=\"nofollow\">https://cvlab-kaist.github.io/GARD/</a></p>\n","updatedAt":"2026-05-27T01:38:03.817Z","author":{"_id":"65ec3449a69aaabb431db0da","avatarUrl":"/avatars/d7b507be0175a61a8fc21176eea45001.svg","fullname":"Jin Hyeon Kim","name":"jinlovespho","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.30657926201820374},"editors":["jinlovespho"],"editorAvatarUrls":["/avatars/d7b507be0175a61a8fc21176eea45001.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2605.26230","authors":[{"_id":"6a164a96e9aa3c8e322db2a0","user":{"_id":"65ec3449a69aaabb431db0da","avatarUrl":"/avatars/d7b507be0175a61a8fc21176eea45001.svg","isPro":false,"fullname":"Jin Hyeon Kim","user":"jinlovespho","type":"user","name":"jinlovespho"},"name":"Jin Hyeon Kim","status":"claimed_verified","statusLastChangedAt":"2026-05-27T07:42:31.211Z","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a1","user":{"_id":"644be3e922d211df644416e9","avatarUrl":"/avatars/6bebbda7a7c16992eba64dc489eaeca5.svg","isPro":false,"fullname":"Jaeeun Lee","user":"alicia10","type":"user","name":"alicia10"},"name":"Jaeeun Lee","status":"claimed_verified","statusLastChangedAt":"2026-05-27T07:53:24.937Z","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a2","name":"Claire Kim","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a3","name":"Kyoungjin Oh","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a4","name":"Paul Hyunbin Cho","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a5","name":"Jaewon Min","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a6","name":"Yeji Choi","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a7","name":"Jihye Park","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a8","name":"Hyunhee Park","hidden":false},{"_id":"6a164a96e9aa3c8e322db2a9","name":"Minkyu Park","hidden":false},{"_id":"6a164a96e9aa3c8e322db2aa","name":"Seungryong Kim","hidden":false}],"publishedAt":"2026-05-25T00:00:00.000Z","submittedOnDailyAt":"2026-05-27T00:00:00.000Z","title":"Geometry-Aware Representation Denoising for Robust Multi-view 3D Reconstruction","submittedOnDailyBy":{"_id":"65ec3449a69aaabb431db0da","avatarUrl":"/avatars/d7b507be0175a61a8fc21176eea45001.svg","isPro":false,"fullname":"Jin Hyeon Kim","user":"jinlovespho","type":"user","name":"jinlovespho"},"summary":"Multi-view 3D reconstruction has achieved remarkable progress with the advent of feed-forward 3D reconstruction models. However, these models are typically trained and evaluated under ideal, degradation-free imaging conditions, whereas real-world observations often contain degradations that differ significantly from such settings. Improving robustness for multi-view 3D reconstruction under degraded conditions therefore remains an important challenge. We present Geometry-Aware Representation Denoising (GARD), a novel framework that performs diffusion-based multi-view restoration directly in the feature space of a feed-forward 3D reconstruction model. This design exploits the geometry-aware feature representations of the 3D reconstructor to effectively recover accurate scene geometry. Furthermore, by employing an additional RGB image decoder, the refined representations can also be used to restore high-quality RGB images, thereby enabling the simultaneous recovery of 3D scene geometry and high-quality imagery. Comprehensive experiments on the Depth Anything 3 (DA3) benchmark demonstrate the effectiveness of the proposed GARD framework.","upvotes":32,"discussionId":"6a164a96e9aa3c8e322db2ab","projectPage":"https://cvlab-kaist.github.io/GARD/","githubRepo":"https://github.com/cvlab-kaist/GARD","githubRepoAddedBy":"user","ai_summary":"A novel diffusion-based framework for multi-view 3D reconstruction that restores both scene geometry and high-quality imagery from degraded inputs by operating in the feature space of a 3D reconstructor.","ai_keywords":["multi-view 3D reconstruction","diffusion-based restoration","feature space","geometry-aware representations","RGB image decoder","Depth Anything 3 benchmark"],"githubStars":10,"organization":{"_id":"6475760c33192631bad2bb38","name":"kaist-ai","fullname":"KAIST AI","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/6469949654873f0043b09c22/aaZFiyXe1qR-Dmy_xq67m.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"65ec3449a69aaabb431db0da","avatarUrl":"/avatars/d7b507be0175a61a8fc21176eea45001.svg","isPro":false,"fullname":"Jin Hyeon Kim","user":"jinlovespho","type":"user"},{"_id":"6752b6315281c3cae4b0783f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/xmcyVEl2xBhk3G5_7dmpz.png","isPro":false,"fullname":"Paul Hyunbin Cho","user":"paulcho98","type":"user"},{"_id":"644be3e922d211df644416e9","avatarUrl":"/avatars/6bebbda7a7c16992eba64dc489eaeca5.svg","isPro":false,"fullname":"Jaeeun Lee","user":"alicia10","type":"user"},{"_id":"652554ff88514c588fb9ea01","avatarUrl":"/avatars/50f2218632d1423980a3e5bef4e1c4e8.svg","isPro":false,"fullname":"Junghyun Park","user":"jamespark30","type":"user"},{"_id":"67c7b179e3f9241dde9ff772","avatarUrl":"/avatars/37cc7a744d8077a0fe7d926cde9d52b2.svg","isPro":false,"fullname":"LeeJaeho","user":"Jaeho0810","type":"user"},{"_id":"67a0c5d12bb4aaf420c6a9f2","avatarUrl":"/avatars/a3b3d994b02fa06fa8de7af1b25f3cb1.svg","isPro":false,"fullname":"Wooseok Jang","user":"woo1726","type":"user"},{"_id":"69897bc008275c850b65442a","avatarUrl":"/avatars/13c5443cec6edf9910d980b965344c91.svg","isPro":false,"fullname":"Seungho Jang","user":"develop0235","type":"user"},{"_id":"67e3a3cc0c2f0d766d401bdb","avatarUrl":"/avatars/0de4c3b11295505ec9d3626e65302cbd.svg","isPro":false,"fullname":"Siyoon Jin","user":"clwm515","type":"user"},{"_id":"668e33799c9aa124a3c69ce0","avatarUrl":"/avatars/e7efef019e5b0447a8e78f6ba70eb4f4.svg","isPro":true,"fullname":"joungbinlee","user":"joungbinlee","type":"user"},{"_id":"68c2353d2c9041473634de6f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/8SnPa72w19yfv8JTYqQPC.png","isPro":false,"fullname":"Claire Seol Kim","user":"kimko-3386","type":"user"},{"_id":"674e9fd3644d6056e583166c","avatarUrl":"/avatars/31dce6e8c1ff256372889c796c67e651.svg","isPro":false,"fullname":"Jinhyuk Jang","user":"JinhyukJang","type":"user"},{"_id":"66f3f780169b85adf963508d","avatarUrl":"/avatars/495ce69c478c332b5fafe897bf1ee80e.svg","isPro":false,"fullname":"Jaeyeong Kim","user":"jy9394","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"6475760c33192631bad2bb38","name":"kaist-ai","fullname":"KAIST AI","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/6469949654873f0043b09c22/aaZFiyXe1qR-Dmy_xq67m.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2605/2605.26230.md"}">
Geometry-Aware Representation Denoising for Robust Multi-view 3D Reconstruction
Abstract
A novel diffusion-based framework for multi-view 3D reconstruction that restores both scene geometry and high-quality imagery from degraded inputs by operating in the feature space of a 3D reconstructor.
AI-generated summary
Multi-view 3D reconstruction has achieved remarkable progress with the advent of feed-forward 3D reconstruction models. However, these models are typically trained and evaluated under ideal, degradation-free imaging conditions, whereas real-world observations often contain degradations that differ significantly from such settings. Improving robustness for multi-view 3D reconstruction under degraded conditions therefore remains an important challenge. We present Geometry-Aware Representation Denoising (GARD), a novel framework that performs diffusion-based multi-view restoration directly in the feature space of a feed-forward 3D reconstruction model. This design exploits the geometry-aware feature representations of the 3D reconstructor to effectively recover accurate scene geometry. Furthermore, by employing an additional RGB image decoder, the refined representations can also be used to restore high-quality RGB images, thereby enabling the simultaneous recovery of 3D scene geometry and high-quality imagery. Comprehensive experiments on the Depth Anything 3 (DA3) benchmark demonstrate the effectiveness of the proposed GARD framework.
Community
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/2605.26230 in a model README.md to link it from this page.
Cite arxiv.org/abs/2605.26230 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2605.26230 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.