Hugging Face Daily Papers · · 3 min read

ReDesign: Recovering Editable Design Structures from Images via Agentic Decomposition

Mirrored from Hugging Face Daily Papers for archival readability. Support the source by reading on the original site.

Data also available at: <a href=\"https://huggingface.co/datasets/Jintae-Park/ReDesign-Figma909\">https://huggingface.co/datasets/Jintae-Park/ReDesign-Figma909</a></p>\n","updatedAt":"2026-07-29T03:54:56.092Z","author":{"_id":"6369f693bf21b20c5692937b","avatarUrl":"/avatars/e937dc8234b3e456149882bfce34841f.svg","fullname":"Jooyeol Yun","name":"YeolJoo","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5776845812797546},"editors":["YeolJoo"],"editorAvatarUrls":["/avatars/e937dc8234b3e456149882bfce34841f.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2607.25565","authors":[{"_id":"6a6978d39d3a1231d492b8a9","user":{"_id":"6369f693bf21b20c5692937b","avatarUrl":"/avatars/e937dc8234b3e456149882bfce34841f.svg","isPro":false,"fullname":"Jooyeol Yun","user":"YeolJoo","type":"user","name":"YeolJoo"},"name":"Jooyeol Yun","status":"claimed_verified","statusLastChangedAt":"2026-07-29T08:45:04.493Z","hidden":false},{"_id":"6a6978d39d3a1231d492b8aa","name":"Jintae Park","hidden":false},{"_id":"6a6978d39d3a1231d492b8ab","name":"Hyesu Lim","hidden":false},{"_id":"6a6978d39d3a1231d492b8ac","name":"Junha Hyung","hidden":false},{"_id":"6a6978d39d3a1231d492b8ad","name":"Hyungjin Chung","hidden":false},{"_id":"6a6978d39d3a1231d492b8ae","name":"Jaegul Choo","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/6369f693bf21b20c5692937b/DdvUj2HnHJGuCtQbHTDfp.mp4"],"publishedAt":"2026-07-28T00:00:00.000Z","submittedOnDailyAt":"2026-07-29T00:00:00.000Z","title":"ReDesign: Recovering Editable Design Structures from Images via Agentic Decomposition","submittedOnDailyBy":{"_id":"6369f693bf21b20c5692937b","avatarUrl":"/avatars/e937dc8234b3e456149882bfce34841f.svg","isPro":false,"fullname":"Jooyeol Yun","user":"YeolJoo","type":"user","name":"YeolJoo"},"summary":"Recovering an editable design file from a raster image is a common and costly bottleneck in modern design workflows, yet remains challenging since editability depends on recovering multi-modal attributes, such as typography, vector geometry, colors, grouping, and layer ordering. We present ReDesign, an agentic framework that grows an editable layer hierarchy by selecting and composing specialized tools across modalities. To keep this long decision process reliable despite imperfect tool outputs, we introduce graceful verification at each expansion, which provides local accept, prune, or retry feedback that prevents error accumulation and avoids large scale reruns. To evaluate editability at scale, we introduce the Figma Edit Replay Benchmark, consisting of 909 raw Figma files and 14,796 controlled edit instructions that replay edits on reconstructed outputs. Across this benchmark and standard reconstruction metrics, ReDesign achieves strong visual fidelity while delivering the highest editability across layout, color, and text edits, outperforming layered decomposition baselines and serial tool use pipelines.","upvotes":36,"discussionId":"6a6978d49d3a1231d492b8af","projectPage":"https://jintae-00.github.io/ReDesign/","githubRepo":"https://github.com/jintae-00/ReDesign","githubRepoAddedBy":"user","githubStars":0,"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":"6449db44df4e6cb7eaef912a","avatarUrl":"/avatars/777cea252e06933863bda10dc3543f59.svg","isPro":false,"fullname":"Min-Jung Kim","user":"emjay73","type":"user"},{"_id":"64797735a68454566356b708","avatarUrl":"/avatars/3424d022dd8ad29b56eb41814c5c3dee.svg","isPro":false,"fullname":"Kinam Kim","user":"kinam0252","type":"user"},{"_id":"6369f693bf21b20c5692937b","avatarUrl":"/avatars/e937dc8234b3e456149882bfce34841f.svg","isPro":false,"fullname":"Jooyeol Yun","user":"YeolJoo","type":"user"},{"_id":"664df2176bc1025819f81caf","avatarUrl":"/avatars/964e2e4612fe05fa0cbadf6139c63077.svg","isPro":false,"fullname":"taewoongkang","user":"Keh0t0","type":"user"},{"_id":"683eda7d29ab8ea25965be1d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/a3oSBCYbpanTph7aO9Cze.png","isPro":false,"fullname":"rhkgkrwk","user":"rhkgkrwk","type":"user"},{"_id":"683edc4af8a74d44ba565695","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/cFTgvaGrBs67TIAtcYaA2.png","isPro":false,"fullname":"nmnaver","user":"nmnaver","type":"user"},{"_id":"672b178fec84b9c3343a311b","avatarUrl":"/avatars/6659f15e4362eab5393de30aa4a7867e.svg","isPro":false,"fullname":"haley yi","user":"haleyyi","type":"user"},{"_id":"683ed74834bd42741af53ef4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/aWyiSdy6SCiQQWU5qw19A.png","isPro":false,"fullname":"nam-gm","user":"nam-gm","type":"user"},{"_id":"673620886375fff559c584a5","avatarUrl":"/avatars/c27a754cceb6886c21aae52ebf9bd598.svg","isPro":false,"fullname":"Leedohyeon","user":"Leedohyeon","type":"user"},{"_id":"672dc10977a7bc84dc94d112","avatarUrl":"/avatars/361b24554fad4be079f9324dfd2d4d29.svg","isPro":false,"fullname":"JT_Park","user":"Jintae-Park","type":"user"},{"_id":"67061bf00cc840f4462a0705","avatarUrl":"/avatars/0fab53fa981f166483aa8ad61172d32d.svg","isPro":false,"fullname":"Dongyoon","user":"godnpeter","type":"user"},{"_id":"651fc0e9444b79d2c60513bd","avatarUrl":"/avatars/693d535fd99f4e9cedef23442c7686c2.svg","isPro":false,"fullname":"Dohyun Lee","user":"aiclaudev","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":3,"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/2607/2607.25565.md","query":{}}">
Papers
arxiv:2607.25565

ReDesign: Recovering Editable Design Structures from Images via Agentic Decomposition

Published on Jul 28
· Submitted by
Jooyeol Yun
on Jul 29
#3 Paper of the day
Authors:

Abstract

Recovering an editable design file from a raster image is a common and costly bottleneck in modern design workflows, yet remains challenging since editability depends on recovering multi-modal attributes, such as typography, vector geometry, colors, grouping, and layer ordering. We present ReDesign, an agentic framework that grows an editable layer hierarchy by selecting and composing specialized tools across modalities. To keep this long decision process reliable despite imperfect tool outputs, we introduce graceful verification at each expansion, which provides local accept, prune, or retry feedback that prevents error accumulation and avoids large scale reruns. To evaluate editability at scale, we introduce the Figma Edit Replay Benchmark, consisting of 909 raw Figma files and 14,796 controlled edit instructions that replay edits on reconstructed outputs. Across this benchmark and standard reconstruction metrics, ReDesign achieves strong visual fidelity while delivering the highest editability across layout, color, and text edits, outperforming layered decomposition baselines and serial tool use pipelines.

Community

Paper author Paper submitter about 6 hours ago
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images

· Sign up or log in to comment

Get this paper in your agent:

hf papers read 2607.25565
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper

No model linking this paper

Cite arxiv.org/abs/2607.25565 in a model README.md to link it from this page.

Datasets citing this paper

Spaces citing this paper

No Space linking this paper

Cite arxiv.org/abs/2607.25565 in a Space README.md to link it from this page.

Collections including this paper

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.

More from Hugging Face Daily Papers