{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:P3AIX6RFOW5HRMNEPEBOXWCSAH","short_pith_number":"pith:P3AIX6RF","schema_version":"1.0","canonical_sha256":"7ec08bfa2575ba78b1a47902ebd85201f9470c4b613a1e4f0ebed7e3adc36b22","source":{"kind":"arxiv","id":"2412.11077","version":3},"attestation_state":"computed","paper":{"title":"Reason-before-Retrieve: One-Stage Reflective Chain-of-Thoughts for Training-Free Zero-Shot Composed Image Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dongmei Zhang, Gang Xiong, Gaopeng Gou, Jing Yu, Jue Zhang, Qingwei Ling, Qi Wu, Saravan Rajmohan, Xiaoting Qin, Yuanmin Tang","submitted_at":"2024-12-15T06:22:20Z","abstract_excerpt":"Composed Image Retrieval (CIR) aims to retrieve target images that closely resemble a reference image while integrating user-specified textual modifications, thereby capturing user intent more precisely. Existing training-free zero-shot CIR (ZS-CIR) methods often employ a two-stage process: they first generate a caption for the reference image and then use Large Language Models for reasoning to obtain a target description. However, these methods suffer from missing critical visual details and limited reasoning capabilities, leading to suboptimal retrieval performance. To address these challeng"},"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":"2412.11077","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-15T06:22:20Z","cross_cats_sorted":[],"title_canon_sha256":"fae695bd206590629e1ad68d8a54cd9977d5c5e6b0ae934674f20154fd62016b","abstract_canon_sha256":"88f7d9c26af38d70e10698c3a57f87297cbcb0da0ab604346f4ffb385370a151"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:52:19.381578Z","signature_b64":"SZWiyf5vB7TAfIJRrXsevFpLd0kvnMx9ys8JjGoP4whb77ItfS4wJP/+RJERqnTOQ7GWGbrbWSpw47E+FKCECw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7ec08bfa2575ba78b1a47902ebd85201f9470c4b613a1e4f0ebed7e3adc36b22","last_reissued_at":"2026-07-05T09:52:19.381078Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:52:19.381078Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reason-before-Retrieve: One-Stage Reflective Chain-of-Thoughts for Training-Free Zero-Shot Composed Image Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dongmei Zhang, Gang Xiong, Gaopeng Gou, Jing Yu, Jue Zhang, Qingwei Ling, Qi Wu, Saravan Rajmohan, Xiaoting Qin, Yuanmin Tang","submitted_at":"2024-12-15T06:22:20Z","abstract_excerpt":"Composed Image Retrieval (CIR) aims to retrieve target images that closely resemble a reference image while integrating user-specified textual modifications, thereby capturing user intent more precisely. Existing training-free zero-shot CIR (ZS-CIR) methods often employ a two-stage process: they first generate a caption for the reference image and then use Large Language Models for reasoning to obtain a target description. However, these methods suffer from missing critical visual details and limited reasoning capabilities, leading to suboptimal retrieval performance. To address these challeng"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.11077","kind":"arxiv","version":3},"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/2412.11077/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":"2412.11077","created_at":"2026-07-05T09:52:19.381137+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.11077v3","created_at":"2026-07-05T09:52:19.381137+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.11077","created_at":"2026-07-05T09:52:19.381137+00:00"},{"alias_kind":"pith_short_12","alias_value":"P3AIX6RFOW5H","created_at":"2026-07-05T09:52:19.381137+00:00"},{"alias_kind":"pith_short_16","alias_value":"P3AIX6RFOW5HRMNE","created_at":"2026-07-05T09:52:19.381137+00:00"},{"alias_kind":"pith_short_8","alias_value":"P3AIX6RF","created_at":"2026-07-05T09:52:19.381137+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.24634","citing_title":"Resolving Ambiguity in Composed Image Retrieval via Calibrated Interaction","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH","json":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH.json","graph_json":"https://pith.science/api/pith-number/P3AIX6RFOW5HRMNEPEBOXWCSAH/graph.json","events_json":"https://pith.science/api/pith-number/P3AIX6RFOW5HRMNEPEBOXWCSAH/events.json","paper":"https://pith.science/paper/P3AIX6RF"},"agent_actions":{"view_html":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH","download_json":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH.json","view_paper":"https://pith.science/paper/P3AIX6RF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.11077&json=true","fetch_graph":"https://pith.science/api/pith-number/P3AIX6RFOW5HRMNEPEBOXWCSAH/graph.json","fetch_events":"https://pith.science/api/pith-number/P3AIX6RFOW5HRMNEPEBOXWCSAH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH/action/storage_attestation","attest_author":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH/action/author_attestation","sign_citation":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH/action/citation_signature","submit_replication":"https://pith.science/pith/P3AIX6RFOW5HRMNEPEBOXWCSAH/action/replication_record"}},"created_at":"2026-07-05T09:52:19.381137+00:00","updated_at":"2026-07-05T09:52:19.381137+00:00"}