{"paper":{"title":"Towards Accurate Single Panoramic 3D Detection: A Semantic Gaussian Centric Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"PanoGSDet lifts 2D panoramic features into continuous 3D semantic Gaussians for monocular object detection.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Kanglin Ning, Shaoru Sun, Wenrui Li, Xiaopeng Fan, Xingtao Wang, Yiran Zhao","submitted_at":"2026-05-14T09:14:24Z","abstract_excerpt":"Three-dimensional object detection in panoramic imagery is crucial for comprehensive scene understanding, yet accurately mapping 2D features to 3D remains a significant challenge. Prevailing methods often project 2D features onto discrete 3D grids, which break geometric continuity and limit representation efficiency. To overcome this limitation, this paper proposes PanoGSDet, a monocular panoramic 3D detection framework built upon continuous semantic 3D Gaussian representations. The proposed framework comprises a panoramic depth estimation component and a semantic Gaussian component. The panor"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Extensive experiments on the Structured3D dataset demonstrate that our method significantly outperforms existing methods.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That spherical 2D semantic and depth features can be accurately projected and optimized into 3D semantic Gaussians that faithfully represent scene geometry from a single monocular panorama.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PanoGSDet projects panoramic 2D features into optimized semantic 3D Gaussians to generate accurate 3D bounding boxes, outperforming prior methods on the Structured3D dataset.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"PanoGSDet lifts 2D panoramic features into continuous 3D semantic Gaussians for monocular object detection.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ce76fadbc06eedb86cca650f16a99dc73624c4be4b493d8ac90394eb8365ef39"},"source":{"id":"2605.14601","kind":"arxiv","version":1},"verdict":{"id":"4192cf82-9f00-44d3-b985-e5e7109f5148","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T04:56:30.884348Z","strongest_claim":"Extensive experiments on the Structured3D dataset demonstrate that our method significantly outperforms existing methods.","one_line_summary":"PanoGSDet projects panoramic 2D features into optimized semantic 3D Gaussians to generate accurate 3D bounding boxes, outperforming prior methods on the Structured3D dataset.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That spherical 2D semantic and depth features can be accurately projected and optimized into 3D semantic Gaussians that faithfully represent scene geometry from a single monocular panorama.","pith_extraction_headline":"PanoGSDet lifts 2D panoramic features into continuous 3D semantic Gaussians for monocular object detection."},"references":{"count":14,"sample":[{"doi":"","year":2025,"title":"One flight over the gap: A survey from perspective to panoramic vision,","work_id":"bc2a792e-56f8-40e1-82bf-88901a5607fc","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Panoextend: An omnidirectional image super-resolution method based on spherical expansion,","work_id":"089b5d26-ce18-43b4-b66b-c6695fa69c6c","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"3d object detection algorithm for panoramic images with multi-scale convolutional neural network,","work_id":"0889c03e-57f4-44b5-8105-f8c3408496fb","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Eliminating the blind spot: Adapting 3d object detection and monocular depth estimation to 360 panoramic imagery,","work_id":"dfe1dbdb-1faa-41e7-954a-20d153059d53","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"3d object detection from a single fisheye image without a single fisheye training image,","work_id":"1906c381-2eb6-4fce-b1ad-868828c4a994","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":14,"snapshot_sha256":"e7a513a3b5c41ca1f2686f2727d120f63b9299c4dc42835b1afa6ce3d3173ebd","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"e0401acf5b38fb6e4235dc9ea794f1056f8b1d509c11a8cea2838921b62ebbe5"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}