{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:RWN6EI4LZ3HMZYQG4ZWMA4HB7W","short_pith_number":"pith:RWN6EI4L","schema_version":"1.0","canonical_sha256":"8d9be2238bcececce206e66cc070e1fd905bd372435c528d8fda34de72bb7147","source":{"kind":"arxiv","id":"2103.07461","version":1},"attestation_state":"computed","paper":{"title":"Probabilistic two-stage detection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Philipp Kr\\\"ahenb\\\"uhl, Vladlen Koltun, Xingyi Zhou","submitted_at":"2021-03-12T18:56:17Z","abstract_excerpt":"We develop a probabilistic interpretation of two-stage object detection. We show that this probabilistic interpretation motivates a number of common empirical training practices. It also suggests changes to two-stage detection pipelines. Specifically, the first stage should infer proper object-vs-background likelihoods, which should then inform the overall score of the detector. A standard region proposal network (RPN) cannot infer this likelihood sufficiently well, but many one-stage detectors can. We show how to build a probabilistic two-stage detector from any state-of-the-art one-stage det"},"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":"2103.07461","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-03-12T18:56:17Z","cross_cats_sorted":[],"title_canon_sha256":"a0bd48dd855ec9fee5468c25e02affcec26b6baf0508ccc9b353b5c91e5c7400","abstract_canon_sha256":"d1e579e0c5232ae07fb1b7efa4cca8c6deaf9217b2400ef8504b9c09f6d75640"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:22:32.068484Z","signature_b64":"R8HpaNJ+OkpD3UfiF6SGyM343+4eo+bJlJrUPDxzTDfvOfhbkfzfrBPD6UKkV2SAlr+AH11+/20gVBtu7EiBBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8d9be2238bcececce206e66cc070e1fd905bd372435c528d8fda34de72bb7147","last_reissued_at":"2026-07-05T02:22:32.068037Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:22:32.068037Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Probabilistic two-stage detection","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Philipp Kr\\\"ahenb\\\"uhl, Vladlen Koltun, Xingyi Zhou","submitted_at":"2021-03-12T18:56:17Z","abstract_excerpt":"We develop a probabilistic interpretation of two-stage object detection. We show that this probabilistic interpretation motivates a number of common empirical training practices. It also suggests changes to two-stage detection pipelines. Specifically, the first stage should infer proper object-vs-background likelihoods, which should then inform the overall score of the detector. A standard region proposal network (RPN) cannot infer this likelihood sufficiently well, but many one-stage detectors can. We show how to build a probabilistic two-stage detector from any state-of-the-art one-stage det"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.07461","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2103.07461/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"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":"2103.07461","created_at":"2026-07-05T02:22:32.068094+00:00"},{"alias_kind":"arxiv_version","alias_value":"2103.07461v1","created_at":"2026-07-05T02:22:32.068094+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.07461","created_at":"2026-07-05T02:22:32.068094+00:00"},{"alias_kind":"pith_short_12","alias_value":"RWN6EI4LZ3HM","created_at":"2026-07-05T02:22:32.068094+00:00"},{"alias_kind":"pith_short_16","alias_value":"RWN6EI4LZ3HMZYQG","created_at":"2026-07-05T02:22:32.068094+00:00"},{"alias_kind":"pith_short_8","alias_value":"RWN6EI4L","created_at":"2026-07-05T02:22:32.068094+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08201","citing_title":"TMI: Text-to-Image Meets Image-to-Image for Complementary Data Synthesis to Boost Long-Tailed Instance Segmentation","ref_index":55,"is_internal_anchor":true},{"citing_arxiv_id":"2604.27101","citing_title":"A Two Stage Pipeline for Left Atrial Wall Constrained Scar Segmentation and Localization from LGE-MR Images","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23344","citing_title":"Exploring Hierarchical Consistency and Unbiased Objectness for Open-Vocabulary Object Detection","ref_index":40,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W","json":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W.json","graph_json":"https://pith.science/api/pith-number/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/graph.json","events_json":"https://pith.science/api/pith-number/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/events.json","paper":"https://pith.science/paper/RWN6EI4L"},"agent_actions":{"view_html":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W","download_json":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W.json","view_paper":"https://pith.science/paper/RWN6EI4L","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2103.07461&json=true","fetch_graph":"https://pith.science/api/pith-number/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/graph.json","fetch_events":"https://pith.science/api/pith-number/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/action/storage_attestation","attest_author":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/action/author_attestation","sign_citation":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/action/citation_signature","submit_replication":"https://pith.science/pith/RWN6EI4LZ3HMZYQG4ZWMA4HB7W/action/replication_record"}},"created_at":"2026-07-05T02:22:32.068094+00:00","updated_at":"2026-07-05T02:22:32.068094+00:00"}