{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:335CVRWBMV3QW54IXP5RD6PQJI","short_pith_number":"pith:335CVRWB","schema_version":"1.0","canonical_sha256":"defa2ac6c165770b7788bbfb11f9f04a27249cb4bfa543be3a1a60f4901e2416","source":{"kind":"arxiv","id":"2503.02304","version":2},"attestation_state":"computed","paper":{"title":"A Token-level Text Image Foundation Model for Document Understanding","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chen Duan, Hao Sun, Junfeng Luo, Kai Zhou, Pei Fu, Qianyi Jiang, Tiezhu Yue, Tongkun Guan, Wei Shen, Xiaokang Yang, Zhengtao Guo, Zining Wang","submitted_at":"2025-03-04T06:05:33Z","abstract_excerpt":"In recent years, general visual foundation models (VFMs) have witnessed increasing adoption, particularly as image encoders for popular multi-modal large language models (MLLMs). However, without semantically fine-grained supervision, these models still encounter fundamental prediction errors in the context of downstream text-image-related tasks, i.e., perception, understanding and reasoning with images containing small and dense texts. To bridge this gap, we develop TokenOCR, the first token-level visual foundation model specifically tailored for text-image-related tasks, designed to support "},"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":"2503.02304","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-04T06:05:33Z","cross_cats_sorted":[],"title_canon_sha256":"3de66b3f9d62aa3b79abde15dee4fe3a30d6cbc1484dd7a43abbb8b39defb896","abstract_canon_sha256":"0ab4ae1c6a13dc41d067bec6282d4b288697b6607433b9ee547c34bcd4cc49f7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:31:55.282668Z","signature_b64":"MfZlzBWpwqFfToWNmAzbg5ZX6EARkZgYiKNKyqmg8Ahe0CQUj+GmfOhtyXofC/Rv4EVA9Ws+bXn9PFvqr9QtAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"defa2ac6c165770b7788bbfb11f9f04a27249cb4bfa543be3a1a60f4901e2416","last_reissued_at":"2026-07-05T10:31:55.282051Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:31:55.282051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Token-level Text Image Foundation Model for Document Understanding","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chen Duan, Hao Sun, Junfeng Luo, Kai Zhou, Pei Fu, Qianyi Jiang, Tiezhu Yue, Tongkun Guan, Wei Shen, Xiaokang Yang, Zhengtao Guo, Zining Wang","submitted_at":"2025-03-04T06:05:33Z","abstract_excerpt":"In recent years, general visual foundation models (VFMs) have witnessed increasing adoption, particularly as image encoders for popular multi-modal large language models (MLLMs). However, without semantically fine-grained supervision, these models still encounter fundamental prediction errors in the context of downstream text-image-related tasks, i.e., perception, understanding and reasoning with images containing small and dense texts. To bridge this gap, we develop TokenOCR, the first token-level visual foundation model specifically tailored for text-image-related tasks, designed to support "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.02304","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2503.02304/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":"2503.02304","created_at":"2026-07-05T10:31:55.282120+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.02304v2","created_at":"2026-07-05T10:31:55.282120+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.02304","created_at":"2026-07-05T10:31:55.282120+00:00"},{"alias_kind":"pith_short_12","alias_value":"335CVRWBMV3Q","created_at":"2026-07-05T10:31:55.282120+00:00"},{"alias_kind":"pith_short_16","alias_value":"335CVRWBMV3QW54I","created_at":"2026-07-05T10:31:55.282120+00:00"},{"alias_kind":"pith_short_8","alias_value":"335CVRWB","created_at":"2026-07-05T10:31:55.282120+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17270","citing_title":"Beyond Detection: A Structure-Aware Framework for Scene Text Tracking","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14708","citing_title":"StyleTextGen: Style-Conditioned Multilingual Scene Text Generation","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI","json":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI.json","graph_json":"https://pith.science/api/pith-number/335CVRWBMV3QW54IXP5RD6PQJI/graph.json","events_json":"https://pith.science/api/pith-number/335CVRWBMV3QW54IXP5RD6PQJI/events.json","paper":"https://pith.science/paper/335CVRWB"},"agent_actions":{"view_html":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI","download_json":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI.json","view_paper":"https://pith.science/paper/335CVRWB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.02304&json=true","fetch_graph":"https://pith.science/api/pith-number/335CVRWBMV3QW54IXP5RD6PQJI/graph.json","fetch_events":"https://pith.science/api/pith-number/335CVRWBMV3QW54IXP5RD6PQJI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI/action/storage_attestation","attest_author":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI/action/author_attestation","sign_citation":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI/action/citation_signature","submit_replication":"https://pith.science/pith/335CVRWBMV3QW54IXP5RD6PQJI/action/replication_record"}},"created_at":"2026-07-05T10:31:55.282120+00:00","updated_at":"2026-07-05T10:31:55.282120+00:00"}