{"paper":{"title":"PanoPlane: Plane-Aware Panoramic Completion for Sparse-View Indoor 3D Gaussian Splatting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"PanoPlane achieves high-fidelity indoor novel view synthesis from sparse inputs by using plane-aware panoramic completion to supervise 3D Gaussian Splatting.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Adil Qureshi, Dinesh Manocha, Dongki Jung, Jaehoon Choi","submitted_at":"2026-05-13T21:39:01Z","abstract_excerpt":"We present PanoPlane, an approach for high-fidelity sparse-view indoor novel view synthesis that reconstructs closed room geometry via panoramic scene completion. Unlike perspective-based methods that generate training views from limited fields of view, PanoPlane leverages $360^{\\circ}$ panoramic completion to condition the generative process on the full spatial layout. We propose Layout Anchored Attention Steering, a training-free mechanism that steers attention within the diffusion model's internal representation toward scene's detected planar surfaces at inference time. By directing each un"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on Replica, ScanNet++, and Matterport3D demonstrate state-of-the-art novel view synthesis quality across 3, 6, and 9 input views, achieving up to +17.8% improvement in PSNR over the current state-of-the-art baseline without any training or fine-tuning of the diffusion model.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That steering attention in the diffusion model toward detected planar surfaces at inference time will reliably produce geometrically consistent extrapolations in unobserved regions without artifacts or inconsistencies.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PanoPlane achieves up to 17.8% PSNR gains in sparse-view indoor novel view synthesis by using training-free plane-aware panoramic completion to supervise 3D Gaussian Splatting.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PanoPlane achieves high-fidelity indoor novel view synthesis from sparse inputs by using plane-aware panoramic completion to supervise 3D Gaussian Splatting.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"7d101307e96db89204972e9777b68ee10ccbbc92461a872576f0d1ef683edbcd"},"source":{"id":"2605.14135","kind":"arxiv","version":1},"verdict":{"id":"2eea33b4-f11b-41cf-b853-051d3cbf4278","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T05:00:27.260109Z","strongest_claim":"Experiments on Replica, ScanNet++, and Matterport3D demonstrate state-of-the-art novel view synthesis quality across 3, 6, and 9 input views, achieving up to +17.8% improvement in PSNR over the current state-of-the-art baseline without any training or fine-tuning of the diffusion model.","one_line_summary":"PanoPlane achieves up to 17.8% PSNR gains in sparse-view indoor novel view synthesis by using training-free plane-aware panoramic completion to supervise 3D Gaussian Splatting.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That steering attention in the diffusion model toward detected planar surfaces at inference time will reliably produce geometrically consistent extrapolations in unobserved regions without artifacts or inconsistencies.","pith_extraction_headline":"PanoPlane achieves high-fidelity indoor novel view synthesis from sparse inputs by using plane-aware panoramic completion to supervise 3D Gaussian Splatting."},"references":{"count":71,"sample":[{"doi":"10.1007/978-3-031-73464-9_1","year":2024,"title":"Self-rectifying diffusion sampling with perturbed-attention guidance","work_id":"39609044-f153-485c-9323-79c321d2a3be","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","ref_index":2,"cited_arxiv_id":"2511.21631","is_internal_anchor":true},{"doi":"","year":2023,"title":"Masactrl: Tuning-free mutual self-attention control for consistent image synthesis and editing","work_id":"3c7b123b-032a-4f01-98d1-f13abe873351","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"Matterport3d: Learning from rgb-d data in indoor environments","work_id":"a6675134-1bd7-4d3f-9344-d7072e7449e9","ref_index":4,"cited_arxiv_id":"1709.06158","is_internal_anchor":true},{"doi":"","year":2025,"title":"Quantifying and alleviating co-adaptation in sparse-view 3d gaussian splatting","work_id":"06848746-c562-4d93-9133-24573435baea","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":71,"snapshot_sha256":"785cce2e9517a6516d4ba2416fb38a6764dccb71d259b05056160641efd127c2","internal_anchors":5},"formal_canon":{"evidence_count":2,"snapshot_sha256":"026e9eab126bc057f883dafbfdc4fe9a7d344068b88033735b099ff699fa2f6d"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}