{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:QOZ6HVA2SLZMH4SBIVEYYS2OEV","short_pith_number":"pith:QOZ6HVA2","schema_version":"1.0","canonical_sha256":"83b3e3d41a92f2c3f24145498c4b4e2547fa859c24ad0ea0b06a2b26397a67ae","source":{"kind":"arxiv","id":"1902.09630","version":2},"attestation_state":"computed","paper":{"title":"Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Amir Sadeghian, Hamid Rezatofighi, Ian Reid, JunYoung Gwak, Nathan Tsoi, Silvio Savarese","submitted_at":"2019-02-25T21:47:33Z","abstract_excerpt":"Intersection over Union (IoU) is the most popular evaluation metric used in the object detection benchmarks. However, there is a gap between optimizing the commonly used distance losses for regressing the parameters of a bounding box and maximizing this metric value. The optimal objective for a metric is the metric itself. In the case of axis-aligned 2D bounding boxes, it can be shown that $IoU$ can be directly used as a regression loss. However, $IoU$ has a plateau making it infeasible to optimize in the case of non-overlapping bounding boxes. In this paper, we address the weaknesses of $IoU$"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"1902.09630","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2019-02-25T21:47:33Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"c76e8b79dee2a4b3292cfbd1b6ffbda4a6d480d26853ba79fe310b572b882512","abstract_canon_sha256":"72760d56da7fede1319d68372c001e11aba78643548068520ebc7a0ae8535667"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:48:39.451439Z","signature_b64":"OiTnbWqdJFotCldSZUBH9wTl8U+jVDaXT/fIm2/RdrLe3tkZAxk762ROZTmRwaGHgb4HIrnm5fOxTGFzwhuXBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"83b3e3d41a92f2c3f24145498c4b4e2547fa859c24ad0ea0b06a2b26397a67ae","last_reissued_at":"2026-05-17T23:48:39.450946Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:48:39.450946Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Amir Sadeghian, Hamid Rezatofighi, Ian Reid, JunYoung Gwak, Nathan Tsoi, Silvio Savarese","submitted_at":"2019-02-25T21:47:33Z","abstract_excerpt":"Intersection over Union (IoU) is the most popular evaluation metric used in the object detection benchmarks. However, there is a gap between optimizing the commonly used distance losses for regressing the parameters of a bounding box and maximizing this metric value. The optimal objective for a metric is the metric itself. In the case of axis-aligned 2D bounding boxes, it can be shown that $IoU$ can be directly used as a regression loss. However, $IoU$ has a plateau making it infeasible to optimize in the case of non-overlapping bounding boxes. In this paper, we address the weaknesses of $IoU$"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1902.09630","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"1902.09630","created_at":"2026-05-17T23:48:39.451018+00:00"},{"alias_kind":"arxiv_version","alias_value":"1902.09630v2","created_at":"2026-05-17T23:48:39.451018+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1902.09630","created_at":"2026-05-17T23:48:39.451018+00:00"},{"alias_kind":"pith_short_12","alias_value":"QOZ6HVA2SLZM","created_at":"2026-05-18T12:33:27.125529+00:00"},{"alias_kind":"pith_short_16","alias_value":"QOZ6HVA2SLZMH4SB","created_at":"2026-05-18T12:33:27.125529+00:00"},{"alias_kind":"pith_short_8","alias_value":"QOZ6HVA2","created_at":"2026-05-18T12:33:27.125529+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.01563","citing_title":"Multi-Dataset Cross-Domain Knowledge Distillation for Unified Medical Image Segmentation, Classification, and Detection","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08626","citing_title":"WildDet3D: Scaling Promptable 3D Detection in the Wild","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV","json":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV.json","graph_json":"https://pith.science/api/pith-number/QOZ6HVA2SLZMH4SBIVEYYS2OEV/graph.json","events_json":"https://pith.science/api/pith-number/QOZ6HVA2SLZMH4SBIVEYYS2OEV/events.json","paper":"https://pith.science/paper/QOZ6HVA2"},"agent_actions":{"view_html":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV","download_json":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV.json","view_paper":"https://pith.science/paper/QOZ6HVA2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1902.09630&json=true","fetch_graph":"https://pith.science/api/pith-number/QOZ6HVA2SLZMH4SBIVEYYS2OEV/graph.json","fetch_events":"https://pith.science/api/pith-number/QOZ6HVA2SLZMH4SBIVEYYS2OEV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV/action/storage_attestation","attest_author":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV/action/author_attestation","sign_citation":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV/action/citation_signature","submit_replication":"https://pith.science/pith/QOZ6HVA2SLZMH4SBIVEYYS2OEV/action/replication_record"}},"created_at":"2026-05-17T23:48:39.451018+00:00","updated_at":"2026-05-17T23:48:39.451018+00:00"}