Hugging Face Daily Papers · August 5, 2026 · 3 min read

UniWorld-Design: From Pixel Generation to Layer-Native Design

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

Like Read original ↗

Paper: <a href=\"https://arxiv.org/abs/2608.03971\" rel=\"nofollow\">https://arxiv.org/abs/2608.03971</a></p>\n","updatedAt":"2026-08-05T02:13:43.262Z","author":{"_id":"646df3c04ad7f907279f14c3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/646df3c04ad7f907279f14c3/WZjSDtAmezmjbczLtCP2B.jpeg","fullname":"Zongjian Li","name":"chestnutlzj","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":30,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.5135867595672607},"editors":["chestnutlzj"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/646df3c04ad7f907279f14c3/WZjSDtAmezmjbczLtCP2B.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.03971","authors":[{"_id":"6a729bba1a375f948521c38e","name":"Zongjian Li","hidden":false},{"_id":"6a729bba1a375f948521c38f","name":"Zhiyuan Yan","hidden":false},{"_id":"6a729bba1a375f948521c390","name":"Chenxu Bai","hidden":false},{"_id":"6a729bba1a375f948521c391","name":"Chen Chen","hidden":false},{"_id":"6a729bba1a375f948521c392","name":"Haoxiang Sun","hidden":false},{"_id":"6a729bba1a375f948521c393","name":"Shaodong Wang","hidden":false},{"_id":"6a729bba1a375f948521c394","name":"Feize Wu","hidden":false},{"_id":"6a729bba1a375f948521c395","name":"Shenghai Yuan","hidden":false},{"_id":"6a729bba1a375f948521c396","name":"Bin Lin","hidden":false},{"_id":"6a729bba1a375f948521c397","name":"Zheyuan Liu","hidden":false},{"_id":"6a729bba1a375f948521c398","name":"Yuwei Niu","hidden":false},{"_id":"6a729bba1a375f948521c399","name":"Li Yuan","hidden":false}],"publishedAt":"2026-08-04T00:00:00.000Z","submittedOnDailyAt":"2026-08-05T00:00:00.000Z","title":"UniWorld-Design: From Pixel Generation to Layer-Native Design","submittedOnDailyBy":{"_id":"646df3c04ad7f907279f14c3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/646df3c04ad7f907279f14c3/WZjSDtAmezmjbczLtCP2B.jpeg","isPro":false,"fullname":"Zongjian Li","user":"chestnutlzj","type":"user","name":"chestnutlzj"},"summary":"We introduce UniWorld-Design, a framework that redefines image generation from flat pixel synthesis to structured visual composition, with semantic RGBA layers as the atomic units of generation, understanding, and editing. Our key insight is that pixels define how an image is rendered, whereas layers define how an image is created, understood, and edited. Just as human designers create and manipulate visual content through layers rather than raw pixels, UniWorld-Design equips multimodal generative models with a layer-native design space. UniWorld-Design comprises two models. The Text-to-RGBA (T2RGBA) model generates standalone RGBA assets directly from text. The Image-to-Layer (I2L) model conditions on a finished image, a global instruction and per-layer prompts, and jointly produces ordered, complete semantic RGBA layers. Its instruction interface supports top-level decomposition, recursive decomposition and targeted extraction, making layering an instruction-addressable operation for agentic editing. Because I2L learns complete semantic objects rather than visible-pixel partitions, its layers stay usable when moved or removed. On the Crello benchmark, I2L reduces per-layer RGB L1 error by 37% and achieves a 34% relative improvement in Alpha Soft IoU over Qwen-Image-Layered. Separately, T2RGBA achieves the highest CLIP Score, outperforming LayerDiffuse and OmniAlpha.","upvotes":15,"discussionId":"6a729bba1a375f948521c39a","projectPage":"https://rabbitvis.rabbitpre.com/","organization":{"_id":"61dcd8e344f59573371b5cb6","name":"PekingUniversity","fullname":"Peking University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/vavgrBsnkSejriUF4lXDE.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"646df3c04ad7f907279f14c3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/646df3c04ad7f907279f14c3/WZjSDtAmezmjbczLtCP2B.jpeg","isPro":false,"fullname":"Zongjian Li","user":"chestnutlzj","type":"user"},{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},{"_id":"63468720dd6d90d82ccf3450","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63468720dd6d90d82ccf3450/tVBFlmZNz8FRMkOrDaDID.jpeg","isPro":false,"fullname":"YSH","user":"BestWishYsh","type":"user"},{"_id":"6463554dd2044cd1d7c6e0bf","avatarUrl":"/avatars/d7653623117268c545a7063fec69664b.svg","isPro":false,"fullname":"Bingzheng Wei","user":"Bingzheng","type":"user"},{"_id":"68f59ae49315a06ad9a01464","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/68f59ae49315a06ad9a01464/welpgtUr6TY9Qa7nPBigg.jpeg","isPro":false,"fullname":"Sean Yu","user":"yushaohan","type":"user"},{"_id":"6a69eb859ef85244dc4ca374","avatarUrl":"/avatars/21fc260ba33de7f03d663e84fba9199b.svg","isPro":false,"fullname":"Anthony Sanchez","user":"anthony-code","type":"user"},{"_id":"6a6a9b15ba280f1955230f14","avatarUrl":"/avatars/a5b5a109e7bbccac09e88a81872662aa.svg","isPro":false,"fullname":"Edward Williams","user":"ZenithMap","type":"user"},{"_id":"6a6c7b702b8f6bcb1b61caba","avatarUrl":"/avatars/40259c5b4eb7b81d54b23737398a112f.svg","isPro":false,"fullname":"Sarah Clark","user":"sarah-clark","type":"user"},{"_id":"6a6c8836e34d1f023f2bfc9b","avatarUrl":"/avatars/b100d4c55e4c331a0360b8338388f349.svg","isPro":false,"fullname":"Anthony Gonzalez","user":"Anthony-Gonzalez","type":"user"},{"_id":"6a6c8c10c9a43ea10742deee","avatarUrl":"/avatars/9b3742550832094a81800eb88787eee8.svg","isPro":false,"fullname":"Joshua Anderson","user":"graniteGlade","type":"user"},{"_id":"6a6dc899c2a26ede038b75c7","avatarUrl":"/avatars/1466d0905927c19ed4470fff1032a0db.svg","isPro":false,"fullname":"Robert Gonzalez","user":"kevin-9736718","type":"user"},{"_id":"6a6dc9e9097052be15666854","avatarUrl":"/avatars/a924e428da37f239c4bdf89900e1ba35.svg","isPro":false,"fullname":"Patricia Gonzalez","user":"michael-1257331","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"organization":{"_id":"61dcd8e344f59573371b5cb6","name":"PekingUniversity","fullname":"Peking University","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/vavgrBsnkSejriUF4lXDE.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.03971.md","query":{}}">

Papers

arxiv:2608.03971

UniWorld-Design: From Pixel Generation to Layer-Native Design

Published on Aug 4

· Submitted by

Zongjian Li on Aug 5

Peking University

Upvote

Authors:

Abstract

We introduce UniWorld-Design, a framework that redefines image generation from flat pixel synthesis to structured visual composition, with semantic RGBA layers as the atomic units of generation, understanding, and editing. Our key insight is that pixels define how an image is rendered, whereas layers define how an image is created, understood, and edited. Just as human designers create and manipulate visual content through layers rather than raw pixels, UniWorld-Design equips multimodal generative models with a layer-native design space. UniWorld-Design comprises two models. The Text-to-RGBA (T2RGBA) model generates standalone RGBA assets directly from text. The Image-to-Layer (I2L) model conditions on a finished image, a global instruction and per-layer prompts, and jointly produces ordered, complete semantic RGBA layers. Its instruction interface supports top-level decomposition, recursive decomposition and targeted extraction, making layering an instruction-addressable operation for agentic editing. Because I2L learns complete semantic objects rather than visible-pixel partitions, its layers stay usable when moved or removed. On the Crello benchmark, I2L reduces per-layer RGB L1 error by 37% and achieves a 34% relative improvement in Alpha Soft IoU over Qwen-Image-Layered. Separately, T2RGBA achieves the highest CLIP Score, outperforming LayerDiffuse and OmniAlpha.

View arXiv page View PDF Project page Add to collection