{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SYCFQ3R2JR3TXFUOO25ILWCIVS","short_pith_number":"pith:SYCFQ3R2","schema_version":"1.0","canonical_sha256":"9604586e3a4c773b968e76ba85d848ac867ea90b14ead57150c35e8ae0b35b36","source":{"kind":"arxiv","id":"2505.05071","version":3},"attestation_state":"computed","paper":{"title":"FG-CLIP: Fine-Grained Visual and Textual Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bin Wang, Chunyu Xie, Dawei Leng, Dawei Liang, Fanjing Kong, Gengshen Zhang, Jincheng Li, Yuhui Yin","submitted_at":"2025-05-08T09:06:53Z","abstract_excerpt":"Contrastive Language-Image Pre-training (CLIP) excels in multimodal tasks such as image-text retrieval and zero-shot classification but struggles with fine-grained understanding due to its focus on coarse-grained short captions. To address this, we propose Fine-Grained CLIP (FG-CLIP), which enhances fine-grained understanding through three key innovations. First, we leverage large multimodal models to generate 1.6 billion long caption-image pairs for capturing global-level semantic details. Second, a high-quality dataset is constructed with 12 million images and 40 million region-specific boun"},"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":"2505.05071","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-05-08T09:06:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c958546b5e03d473346c4843418cc12fd5c685e49d122bf2496346919385c3a7","abstract_canon_sha256":"095e3fede78e2ecfb69e8cc4400784b2fed13189c27931a3c6504c0d650899d0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:31.822975Z","signature_b64":"BtyHSuJoPxiMXTjGvq+feKd8zQgoVs4dvk4CcjdqJUXDQVttnQ7H1GgdRHQ9C4+/LsiH5hhsDkYf7yeBApYlAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9604586e3a4c773b968e76ba85d848ac867ea90b14ead57150c35e8ae0b35b36","last_reissued_at":"2026-07-05T11:06:31.822508Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:31.822508Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FG-CLIP: Fine-Grained Visual and Textual Alignment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bin Wang, Chunyu Xie, Dawei Leng, Dawei Liang, Fanjing Kong, Gengshen Zhang, Jincheng Li, Yuhui Yin","submitted_at":"2025-05-08T09:06:53Z","abstract_excerpt":"Contrastive Language-Image Pre-training (CLIP) excels in multimodal tasks such as image-text retrieval and zero-shot classification but struggles with fine-grained understanding due to its focus on coarse-grained short captions. To address this, we propose Fine-Grained CLIP (FG-CLIP), which enhances fine-grained understanding through three key innovations. First, we leverage large multimodal models to generate 1.6 billion long caption-image pairs for capturing global-level semantic details. Second, a high-quality dataset is constructed with 12 million images and 40 million region-specific boun"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.05071","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/2505.05071/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":"2505.05071","created_at":"2026-07-05T11:06:31.822568+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.05071v3","created_at":"2026-07-05T11:06:31.822568+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.05071","created_at":"2026-07-05T11:06:31.822568+00:00"},{"alias_kind":"pith_short_12","alias_value":"SYCFQ3R2JR3T","created_at":"2026-07-05T11:06:31.822568+00:00"},{"alias_kind":"pith_short_16","alias_value":"SYCFQ3R2JR3TXFUO","created_at":"2026-07-05T11:06:31.822568+00:00"},{"alias_kind":"pith_short_8","alias_value":"SYCFQ3R2","created_at":"2026-07-05T11:06:31.822568+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26794","citing_title":"ReasonCLIP-58M: Visually Grounded Commonsense Reasoning Supervision for CLIP","ref_index":95,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22158","citing_title":"Improving Reasoning in Vision-Language Models via Perception Verified Self-Training","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06978","citing_title":"CL-CLIP: CLIP-Based Continual Learning Framework with Cost-Volume Category Decoupling for Object Detection","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31924","citing_title":"InstanceControl: Controllable Complex Image Generation without Instance Labeling","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22158","citing_title":"Improving Reasoning in Vision-Language Models via Perception Verified Self-Training","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2510.26583","citing_title":"Emu3.5: Native Multimodal Models are World Learners","ref_index":112,"is_internal_anchor":false},{"citing_arxiv_id":"2511.13415","citing_title":"Attention Grounded Enhancement for Visual Document Retrieval","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2511.16567","citing_title":"POMA-3D: The Point Map Way to 3D Scene Understanding","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02546","citing_title":"RGB-Pointmap Pretraining for Unified 3D Scene Understanding","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12013","citing_title":"L2P: Unlocking Latent Potential for Pixel Generation","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08156","citing_title":"LAGO: Language-Guided Adaptive Object-Region Focus for Zero-Shot Visual-Text Alignment","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00526","citing_title":"IdentiFace: Multi-Modal Iterative Diffusion Framework for Identifiable Suspect Face Generation in Crime Investigations","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17982","citing_title":"Mitigating Multimodal Hallucination via Phase-wise Self-reward","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20135","citing_title":"AFMRL: Attribute-Enhanced Fine-Grained Multi-Modal Representation Learning in E-commerce","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS","json":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS.json","graph_json":"https://pith.science/api/pith-number/SYCFQ3R2JR3TXFUOO25ILWCIVS/graph.json","events_json":"https://pith.science/api/pith-number/SYCFQ3R2JR3TXFUOO25ILWCIVS/events.json","paper":"https://pith.science/paper/SYCFQ3R2"},"agent_actions":{"view_html":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS","download_json":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS.json","view_paper":"https://pith.science/paper/SYCFQ3R2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.05071&json=true","fetch_graph":"https://pith.science/api/pith-number/SYCFQ3R2JR3TXFUOO25ILWCIVS/graph.json","fetch_events":"https://pith.science/api/pith-number/SYCFQ3R2JR3TXFUOO25ILWCIVS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS/action/storage_attestation","attest_author":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS/action/author_attestation","sign_citation":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS/action/citation_signature","submit_replication":"https://pith.science/pith/SYCFQ3R2JR3TXFUOO25ILWCIVS/action/replication_record"}},"created_at":"2026-07-05T11:06:31.822568+00:00","updated_at":"2026-07-05T11:06:31.822568+00:00"}