{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:Q56OF553UU77N3YCFSDP7KWWFD","short_pith_number":"pith:Q56OF553","schema_version":"1.0","canonical_sha256":"877ce2f7bba53ff6ef022c86ffaad628c125d243a197d64e833c17f6b93486b6","source":{"kind":"arxiv","id":"2605.18434","version":1},"attestation_state":"computed","paper":{"title":"TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Item text can guide the creation of target-focused representations from full images to match cropped visual queries without object detection.","cross_cats":["cs.CV"],"primary_cat":"cs.IR","authors_text":"Ben Chen, Chenyi Lei, Huangyu Dai, Lingtao Mao, Wenwu Ou, Xinyu Sun, Zexin Zheng, Zihan Liang","submitted_at":"2026-05-18T14:07:20Z","abstract_excerpt":"E-commerce image search often takes a cropped image as the query, while each candidate is represented by full item images and structured text. This image-to-multimodal retrieval setting presents two asymmetries: a modality disparity -- a visual query must match image--text items, and a granularity disparity -- a cropped query must be compared with full images containing background context and possible distractors. Detection-based pipelines handle the granularity disparity through explicit localization but incur extra cost and error propagation, whereas CLIP-style encoders avoid detection, but "},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.18434","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.IR","submitted_at":"2026-05-18T14:07:20Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"34cf35ad5b5106e6537cd79bb700be08f9a01ee018e2daa7d09ff27c569f2527","abstract_canon_sha256":"b270e89a8c91c34614a44460061563da9fe0983b184e0323ab070784f01508f8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:06:00.783448Z","signature_b64":"GAF49ozVGSzYQnE381uX4vAjGRBmMb3WuFAhvWBOBateLFgJJ2jbZ0Bgj6CbLI3+1QKKr6d5Wfzsh5DtCEfuAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"877ce2f7bba53ff6ef022c86ffaad628c125d243a197d64e833c17f6b93486b6","last_reissued_at":"2026-05-20T00:06:00.782578Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:06:00.782578Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Item text can guide the creation of target-focused representations from full images to match cropped visual queries without object detection.","cross_cats":["cs.CV"],"primary_cat":"cs.IR","authors_text":"Ben Chen, Chenyi Lei, Huangyu Dai, Lingtao Mao, Wenwu Ou, Xinyu Sun, Zexin Zheng, Zihan Liang","submitted_at":"2026-05-18T14:07:20Z","abstract_excerpt":"E-commerce image search often takes a cropped image as the query, while each candidate is represented by full item images and structured text. This image-to-multimodal retrieval setting presents two asymmetries: a modality disparity -- a visual query must match image--text items, and a granularity disparity -- a cropped query must be compared with full images containing background context and possible distractors. Detection-based pipelines handle the granularity disparity through explicit localization but incur extra cost and error propagation, whereas CLIP-style encoders avoid detection, but "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"TIGER-FG improves Recall@1 over the strongest baseline by 6.1 and 34.4 percentage points on the two evaluation benchmarks, respectively, with only 85.7M query-side parameters and 256-dim embeddings.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That item text descriptions provide reliable semantic guidance sufficient to produce target-focused representations implicitly, without explicit localization or detection, and that the dual distillation objectives will preserve both spatial consistency and similarity structure across varied item layouts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TIGER-FG proposes text-guided implicit fine-grained grounding with dual distillation to address modality and granularity asymmetries in image-to-multimodal e-commerce retrieval, reporting Recall@1 gains of 6.1 and 34.4 points on two new benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Item text can guide the creation of target-focused representations from full images to match cropped visual queries without object detection.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"54d8b57e21efc56e5578e75491fe339e7f06980f68270cc7d3cc5ff9292ca1f4"},"source":{"id":"2605.18434","kind":"arxiv","version":1},"verdict":{"id":"83dc678e-3ee3-4925-87f2-4e6a01ff738e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T23:51:00.345273Z","strongest_claim":"TIGER-FG improves Recall@1 over the strongest baseline by 6.1 and 34.4 percentage points on the two evaluation benchmarks, respectively, with only 85.7M query-side parameters and 256-dim embeddings.","one_line_summary":"TIGER-FG proposes text-guided implicit fine-grained grounding with dual distillation to address modality and granularity asymmetries in image-to-multimodal e-commerce retrieval, reporting Recall@1 gains of 6.1 and 34.4 points on two new benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That item text descriptions provide reliable semantic guidance sufficient to produce target-focused representations implicitly, without explicit localization or detection, and that the dual distillation objectives will preserve both spatial consistency and similarity structure across varied item layouts.","pith_extraction_headline":"Item text can guide the creation of target-focused representations from full images to match cropped visual queries without object detection."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.18434/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-20T00:01:20.228987Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-20T00:00:55.631652Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"citation_quote_validity","ran_at":"2026-05-19T23:49:47.549594Z","status":"skipped","version":"0.1.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T23:33:27.568210Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"external_links","ran_at":"2026-05-19T23:31:32.146230Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"cited_work_retraction","ran_at":"2026-05-19T23:21:58.699669Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T23:21:58.650004Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"d5fdc3f39f5a86f548feec5608791756ce372f43f7b30a594a1009c97c8c046b"},"references":{"count":19,"sample":[{"doi":"","year":2020,"title":"End-to-end object detection with transformers","work_id":"37ce22dc-54b4-462d-b1bf-6481e8136541","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Ben Chen, Linbo Jin, Xinxin Wang, Dehong Gao, Wen Jiang, and Wei Ning","work_id":"753c4c1a-6490-40d8-bd85-a6539f55d9fd","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Category-oriented representation learning for image to multi-modal retrieval.arXiv preprint arXiv:2305.03972,","work_id":"6dc6f514-1c31-42d7-9454-0bd2091678ee","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Vision Transformers Need Registers","work_id":"57106da4-5420-4778-94eb-e821589aa7a0","ref_index":4,"cited_arxiv_id":"2309.16588","is_internal_anchor":true},{"doi":"","year":null,"title":"The Faiss library","work_id":"41b4c732-3a63-4106-9e7c-50d61cfb2126","ref_index":6,"cited_arxiv_id":"2401.08281","is_internal_anchor":true}],"resolved_work":19,"snapshot_sha256":"7964f028db8911b7afd6c75eeef25bf55851ffb7e0296b4d57d9163fb1baf007","internal_anchors":7},"formal_canon":{"evidence_count":2,"snapshot_sha256":"7a7b89de2f7841591d3427d3ba761a4d25e5da1ae37da8259f677e5619e4fc44"},"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":"2605.18434","created_at":"2026-05-20T00:06:00.782724+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18434v1","created_at":"2026-05-20T00:06:00.782724+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18434","created_at":"2026-05-20T00:06:00.782724+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q56OF553UU77","created_at":"2026-05-20T00:06:00.782724+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q56OF553UU77N3YC","created_at":"2026-05-20T00:06:00.782724+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q56OF553","created_at":"2026-05-20T00:06:00.782724+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD","json":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD.json","graph_json":"https://pith.science/api/pith-number/Q56OF553UU77N3YCFSDP7KWWFD/graph.json","events_json":"https://pith.science/api/pith-number/Q56OF553UU77N3YCFSDP7KWWFD/events.json","paper":"https://pith.science/paper/Q56OF553"},"agent_actions":{"view_html":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD","download_json":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD.json","view_paper":"https://pith.science/paper/Q56OF553","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18434&json=true","fetch_graph":"https://pith.science/api/pith-number/Q56OF553UU77N3YCFSDP7KWWFD/graph.json","fetch_events":"https://pith.science/api/pith-number/Q56OF553UU77N3YCFSDP7KWWFD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD/action/storage_attestation","attest_author":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD/action/author_attestation","sign_citation":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD/action/citation_signature","submit_replication":"https://pith.science/pith/Q56OF553UU77N3YCFSDP7KWWFD/action/replication_record"}},"created_at":"2026-05-20T00:06:00.782724+00:00","updated_at":"2026-05-20T00:06:00.782724+00:00"}