Inspired by classical Bundle Adjustment, we propose BA-T, an iterative Transformer that implements BA-style structured updates as a repeatable layer in implicit token space. Instead of relying on deep attention stacks, BA-T refines predictions based on latent residual by a single lightweight layer.</p>\n","updatedAt":"2026-06-03T12:17:45.302Z","author":{"_id":"6501764b35ec971762b57317","avatarUrl":"/avatars/47d4fa98df0df3d02428b4c3347e239c.svg","fullname":"Ganlin Zhang","name":"zhangganlin","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":1,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8697284460067749},"editors":["zhangganlin"],"editorAvatarUrls":["/avatars/47d4fa98df0df3d02428b4c3347e239c.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.03287","authors":[{"_id":"6a201ab5e292c1c78ecb15ed","name":"Ganlin Zhang","hidden":false},{"_id":"6a201ab5e292c1c78ecb15ee","name":"Weirong Chen","hidden":false},{"_id":"6a201ab5e292c1c78ecb15ef","name":"Daniel Cremers","hidden":false},{"_id":"6a201ab5e292c1c78ecb15f0","name":"Xi Wang","hidden":false}],"publishedAt":"2026-06-02T00:00:00.000Z","submittedOnDailyAt":"2026-06-03T00:00:00.000Z","title":"BA-T: An Iterative Transformer for Two-View Bundle Adjustment","submittedOnDailyBy":{"_id":"6501764b35ec971762b57317","avatarUrl":"/avatars/47d4fa98df0df3d02428b4c3347e239c.svg","isPro":false,"fullname":"Ganlin Zhang","user":"zhangganlin","type":"user","name":"zhangganlin"},"summary":"Feed-forward models for 3D reconstruction have achieved strong performance using deep cross-view attention to exchange information across images. However, these approaches often depend on heavy decoder stacks and lack a structured mechanism for geometry refinement, resulting in poor multi-view consistency. We address this by drawing inspiration from classical bundle adjustment (BA), which can be viewed as an iterative information propagation process between poses and local geometry. Inspired by BA, we propose BA-T, an iterative Transformer that implements BA-style structured updates as a repeatable layer in implicit token space. Instead of relying on deep attention stacks, BA-T refines predictions based on latent residual by a single lightweight layer. Experiments demonstrate that BA-T progressively improves pose and reconstruction accuracy across iterations, achieves stronger cross-view consistency than conventional decoders, and matches or surpasses substantially larger models while using only 16% of their decoder parameters. BA-T provides a compact, efficient, and structural alternative to depth-heavy attention, enabling accurate 3D reconstruction within a lightweight architecture. The code will be made publicly at https://github.com/zhangganlin/BA-T.","upvotes":0,"discussionId":"6a201ab5e292c1c78ecb15f1","githubRepo":"https://github.com/zhangganlin/BA-T","githubRepoAddedBy":"user","ai_summary":"BA-T is an iterative Transformer architecture that improves 3D reconstruction accuracy and cross-view consistency through structured updates inspired by bundle adjustment, using a lightweight design that requires only 16% of conventional decoder parameters.","ai_keywords":["feed-forward models","3D reconstruction","deep cross-view attention","decoder stacks","geometry refinement","bundle adjustment","iterative Transformer","implicit token space","latent residual","cross-view consistency","depth-heavy attention"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":5,"organization":{"_id":"61fae781e68759322b9767be","name":"TUM","fullname":"Technical University of Munich","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1661167219960-629521a0f937190946e15d7f.jpeg"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[],"acceptLanguages":["en"],"organization":{"_id":"61fae781e68759322b9767be","name":"TUM","fullname":"Technical University of Munich","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1661167219960-629521a0f937190946e15d7f.jpeg"}}">
BA-T: An Iterative Transformer for Two-View Bundle Adjustment
Abstract
BA-T is an iterative Transformer architecture that improves 3D reconstruction accuracy and cross-view consistency through structured updates inspired by bundle adjustment, using a lightweight design that requires only 16% of conventional decoder parameters.
Feed-forward models for 3D reconstruction have achieved strong performance using deep cross-view attention to exchange information across images. However, these approaches often depend on heavy decoder stacks and lack a structured mechanism for geometry refinement, resulting in poor multi-view consistency. We address this by drawing inspiration from classical bundle adjustment (BA), which can be viewed as an iterative information propagation process between poses and local geometry. Inspired by BA, we propose BA-T, an iterative Transformer that implements BA-style structured updates as a repeatable layer in implicit token space. Instead of relying on deep attention stacks, BA-T refines predictions based on latent residual by a single lightweight layer. Experiments demonstrate that BA-T progressively improves pose and reconstruction accuracy across iterations, achieves stronger cross-view consistency than conventional decoders, and matches or surpasses substantially larger models while using only 16% of their decoder parameters. BA-T provides a compact, efficient, and structural alternative to depth-heavy attention, enabling accurate 3D reconstruction within a lightweight architecture. The code will be made publicly at https://github.com/zhangganlin/BA-T.
Community
Inspired by classical Bundle Adjustment, we propose BA-T, an iterative Transformer that implements BA-style structured updates as a repeatable layer in implicit token space. Instead of relying on deep attention stacks, BA-T refines predictions based on latent residual by a single lightweight layer.
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/2606.03287 in a model README.md to link it from this page.
Cite arxiv.org/abs/2606.03287 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2606.03287 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.