{"paper":{"title":"MoGe-2: Accurate Monocular Geometry with Metric Scale and Sharp Details","license":"http://creativecommons.org/licenses/by/4.0/","headline":"MoGe-2 recovers metric-scale 3D point maps from single images while preserving relative accuracy and recovering fine details.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Guangzhong Sun, Jianfeng Xiang, Jiaolong Yang, Ruicheng Wang, Sicheng Xu, Xin Tong, Yu Deng, Yue Dong, Zelong Lv","submitted_at":"2025-07-03T11:40:01Z","abstract_excerpt":"We propose MoGe-2, an advanced open-domain geometry estimation model that recovers a metric scale 3D point map of a scene from a single image. Our method builds upon the recent monocular geometry estimation approach, MoGe, which predicts affine-invariant point maps with unknown scales. We explore effective strategies to extend MoGe for metric geometry prediction without compromising the relative geometry accuracy provided by the affine-invariant point representation. Additionally, we discover that noise and errors in real data diminish fine-grained detail in the predicted geometry. We address "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"demonstrating its superior performance in achieving accurate relative geometry, precise metric scale, and fine-grained detail recovery -- capabilities that no previous methods have simultaneously achieved.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That filtering and completing real data with sharp synthetic labels preserves overall accuracy without introducing systematic biases or artifacts in the metric scale prediction.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MoGe-2 recovers metric-scale 3D point maps with fine details from single images via data refinement and extension of affine-invariant predictions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MoGe-2 recovers metric-scale 3D point maps from single images while preserving relative accuracy and recovering fine details.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"c3d3708e42390f7c4efdc12ec35be462a375134cb79ddb30e2c1ac176e31d940"},"source":{"id":"2507.02546","kind":"arxiv","version":1},"verdict":{"id":"9bc5d1f1-9022-4657-a152-bde9404c4c96","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:16:02.251020Z","strongest_claim":"demonstrating its superior performance in achieving accurate relative geometry, precise metric scale, and fine-grained detail recovery -- capabilities that no previous methods have simultaneously achieved.","one_line_summary":"MoGe-2 recovers metric-scale 3D point maps with fine details from single images via data refinement and extension of affine-invariant predictions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That filtering and completing real data with sharp synthetic labels preserves overall accuracy without introducing systematic biases or artifacts in the metric scale prediction.","pith_extraction_headline":"MoGe-2 recovers metric-scale 3D point maps from single images while preserving relative accuracy and recovering fine details."},"references":{"count":83,"sample":[{"doi":"","year":2019,"title":"Apollo synthetic dataset, 2019","work_id":"8864ac7f-ef2c-45b4-8cb3-b1ddde64e709","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Zip-nerf: Anti- aliased grid-based neural radiance fields","work_id":"9639f894-6968-4677-bbb9-2ef05886e3a5","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"ARKitscenes - a diverse real-world dataset for 3d indoor scene understanding using mobile RGB-d data","work_id":"bdc2d9d8-b51a-48e3-afcf-bf862ee08c39","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"Adabins: Depth estimation using adaptive bins","work_id":"4c9b0a43-3a22-4f30-8ab5-f2b1125fbbb9","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"ZoeDepth: Zero-shot Transfer by Combining Relative and Metric Depth","work_id":"c902e427-c54a-4fc4-8aef-d0243f90ea39","ref_index":5,"cited_arxiv_id":"2302.12288","is_internal_anchor":true}],"resolved_work":83,"snapshot_sha256":"780f7627cbc2a2b0ee91654468bc32caaefdebf0c97870c034e8013b257f366c","internal_anchors":7},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4711f00f734e10caad460eaf929240113de1ce56ee645b3fa43300e30537ee54"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}