{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6GCAB457TBMQ6BAGSG6SBMEAOX","short_pith_number":"pith:6GCAB457","schema_version":"1.0","canonical_sha256":"f18400f3bf98590f040691bd20b08075ca6d5f849154e4003ce0cc9b99c481a4","source":{"kind":"arxiv","id":"2504.14032","version":1},"attestation_state":"computed","paper":{"title":"LoftUp: Learning a Coordinate-Based Feature Upsampler for Vision Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","eess.IV"],"primary_cat":"cs.CV","authors_text":"Andreas Geiger, Anpei Chen, Dan Zhang, Haiwen Huang, Volodymyr Havrylov","submitted_at":"2025-04-18T18:46:08Z","abstract_excerpt":"Vision foundation models (VFMs) such as DINOv2 and CLIP have achieved impressive results on various downstream tasks, but their limited feature resolution hampers performance in applications requiring pixel-level understanding. Feature upsampling offers a promising direction to address this challenge. In this work, we identify two critical factors for enhancing feature upsampling: the upsampler architecture and the training objective. For the upsampler architecture, we introduce a coordinate-based cross-attention transformer that integrates the high-resolution images with coordinates and low-r"},"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":"2504.14032","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-04-18T18:46:08Z","cross_cats_sorted":["cs.AI","cs.LG","eess.IV"],"title_canon_sha256":"397bc0af694069326ba484b0add477d567575333ea4eb2a9458c8375ebcc591c","abstract_canon_sha256":"62cede5f5f759f4052069097caa73a8dffea81e64a47599d285d38a55ee209ad"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:51:09.881668Z","signature_b64":"3ZPtAl8AjfF5bNGYQCEWEcUxFmFMGUzFwlLY9ZaaMtmjlLeoqWekJwN5FChpdNCzRXhAPFzrAiYqhlbTby2KCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f18400f3bf98590f040691bd20b08075ca6d5f849154e4003ce0cc9b99c481a4","last_reissued_at":"2026-07-05T10:51:09.881250Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:51:09.881250Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LoftUp: Learning a Coordinate-Based Feature Upsampler for Vision Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","eess.IV"],"primary_cat":"cs.CV","authors_text":"Andreas Geiger, Anpei Chen, Dan Zhang, Haiwen Huang, Volodymyr Havrylov","submitted_at":"2025-04-18T18:46:08Z","abstract_excerpt":"Vision foundation models (VFMs) such as DINOv2 and CLIP have achieved impressive results on various downstream tasks, but their limited feature resolution hampers performance in applications requiring pixel-level understanding. Feature upsampling offers a promising direction to address this challenge. In this work, we identify two critical factors for enhancing feature upsampling: the upsampler architecture and the training objective. For the upsampler architecture, we introduce a coordinate-based cross-attention transformer that integrates the high-resolution images with coordinates and low-r"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.14032","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/2504.14032/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":"2504.14032","created_at":"2026-07-05T10:51:09.881307+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.14032v1","created_at":"2026-07-05T10:51:09.881307+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.14032","created_at":"2026-07-05T10:51:09.881307+00:00"},{"alias_kind":"pith_short_12","alias_value":"6GCAB457TBMQ","created_at":"2026-07-05T10:51:09.881307+00:00"},{"alias_kind":"pith_short_16","alias_value":"6GCAB457TBMQ6BAG","created_at":"2026-07-05T10:51:09.881307+00:00"},{"alias_kind":"pith_short_8","alias_value":"6GCAB457","created_at":"2026-07-05T10:51:09.881307+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX","json":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX.json","graph_json":"https://pith.science/api/pith-number/6GCAB457TBMQ6BAGSG6SBMEAOX/graph.json","events_json":"https://pith.science/api/pith-number/6GCAB457TBMQ6BAGSG6SBMEAOX/events.json","paper":"https://pith.science/paper/6GCAB457"},"agent_actions":{"view_html":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX","download_json":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX.json","view_paper":"https://pith.science/paper/6GCAB457","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.14032&json=true","fetch_graph":"https://pith.science/api/pith-number/6GCAB457TBMQ6BAGSG6SBMEAOX/graph.json","fetch_events":"https://pith.science/api/pith-number/6GCAB457TBMQ6BAGSG6SBMEAOX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX/action/storage_attestation","attest_author":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX/action/author_attestation","sign_citation":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX/action/citation_signature","submit_replication":"https://pith.science/pith/6GCAB457TBMQ6BAGSG6SBMEAOX/action/replication_record"}},"created_at":"2026-07-05T10:51:09.881307+00:00","updated_at":"2026-07-05T10:51:09.881307+00:00"}