{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ICCHL7QYTDJ5B3CSS3TDUP3Z2Z","short_pith_number":"pith:ICCHL7QY","schema_version":"1.0","canonical_sha256":"408475fe1898d3d0ec5296e63a3f79d669d0542f09bbc96f6653e9ac19163158","source":{"kind":"arxiv","id":"2503.21749","version":1},"attestation_state":"computed","paper":{"title":"LeX-Art: Rethinking Text Generation via Scalable High-Quality Data Synthesis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bin Fu, Bo Zhang, Dongyang Liu, Hongsheng Li, Kaipeng Zhang, Ming Li, Peng Gao, Qilong Wu, Qi Qin, Shitian Zhao, Xinyue Li, Yu Qiao, Zhen Li","submitted_at":"2025-03-27T17:56:15Z","abstract_excerpt":"We introduce LeX-Art, a comprehensive suite for high-quality text-image synthesis that systematically bridges the gap between prompt expressiveness and text rendering fidelity. Our approach follows a data-centric paradigm, constructing a high-quality data synthesis pipeline based on Deepseek-R1 to curate LeX-10K, a dataset of 10K high-resolution, aesthetically refined 1024$\\times$1024 images. Beyond dataset construction, we develop LeX-Enhancer, a robust prompt enrichment model, and train two text-to-image models, LeX-FLUX and LeX-Lumina, achieving state-of-the-art text rendering performance. "},"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":"2503.21749","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-03-27T17:56:15Z","cross_cats_sorted":[],"title_canon_sha256":"0f0f83dfe18f82d304eb87feff1ca1b96e07b48842015d72d36614a75c46e728","abstract_canon_sha256":"d1c7badcad6bf32dc4c1eed24a57e8f3d0d97ec9ea2b928e1c7a48a73a382d40"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:40:29.895303Z","signature_b64":"XWqTC9aGFe3cXe31QGCcIWemZJ5n9dZRDQbvwUz5CrmIjRnqm5BK5whKKv7Mkkbw17zBwuTB7HNk0NNYCU8IAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"408475fe1898d3d0ec5296e63a3f79d669d0542f09bbc96f6653e9ac19163158","last_reissued_at":"2026-07-05T10:40:29.894812Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:40:29.894812Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LeX-Art: Rethinking Text Generation via Scalable High-Quality Data Synthesis","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bin Fu, Bo Zhang, Dongyang Liu, Hongsheng Li, Kaipeng Zhang, Ming Li, Peng Gao, Qilong Wu, Qi Qin, Shitian Zhao, Xinyue Li, Yu Qiao, Zhen Li","submitted_at":"2025-03-27T17:56:15Z","abstract_excerpt":"We introduce LeX-Art, a comprehensive suite for high-quality text-image synthesis that systematically bridges the gap between prompt expressiveness and text rendering fidelity. Our approach follows a data-centric paradigm, constructing a high-quality data synthesis pipeline based on Deepseek-R1 to curate LeX-10K, a dataset of 10K high-resolution, aesthetically refined 1024$\\times$1024 images. Beyond dataset construction, we develop LeX-Enhancer, a robust prompt enrichment model, and train two text-to-image models, LeX-FLUX and LeX-Lumina, achieving state-of-the-art text rendering performance. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.21749","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/2503.21749/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":"2503.21749","created_at":"2026-07-05T10:40:29.894869+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.21749v1","created_at":"2026-07-05T10:40:29.894869+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.21749","created_at":"2026-07-05T10:40:29.894869+00:00"},{"alias_kind":"pith_short_12","alias_value":"ICCHL7QYTDJ5","created_at":"2026-07-05T10:40:29.894869+00:00"},{"alias_kind":"pith_short_16","alias_value":"ICCHL7QYTDJ5B3CS","created_at":"2026-07-05T10:40:29.894869+00:00"},{"alias_kind":"pith_short_8","alias_value":"ICCHL7QY","created_at":"2026-07-05T10:40:29.894869+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04479","citing_title":"Evaluating Reasoning Fidelity in Visual Text Generation","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2510.16888","citing_title":"Uniworld-V2: Reinforce Image Editing with Diffusion Negative-aware Finetuning and MLLM Implicit Feedback","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21090","citing_title":"TextSculptor: Training and Benchmarking Scene Text Editing","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2510.26583","citing_title":"Emu3.5: Native Multimodal Models are World Learners","ref_index":129,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08163","citing_title":"MULTITEXTEDIT: Benchmarking Cross-Lingual Degradation in Text-in-Image Editing","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24171","citing_title":"POCA: Pareto-Optimal Curriculum Alignment for Visual Text Generation","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z","json":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z.json","graph_json":"https://pith.science/api/pith-number/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/graph.json","events_json":"https://pith.science/api/pith-number/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/events.json","paper":"https://pith.science/paper/ICCHL7QY"},"agent_actions":{"view_html":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z","download_json":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z.json","view_paper":"https://pith.science/paper/ICCHL7QY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.21749&json=true","fetch_graph":"https://pith.science/api/pith-number/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/graph.json","fetch_events":"https://pith.science/api/pith-number/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/action/storage_attestation","attest_author":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/action/author_attestation","sign_citation":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/action/citation_signature","submit_replication":"https://pith.science/pith/ICCHL7QYTDJ5B3CSS3TDUP3Z2Z/action/replication_record"}},"created_at":"2026-07-05T10:40:29.894869+00:00","updated_at":"2026-07-05T10:40:29.894869+00:00"}