{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:XWQRXIRXOWUZIQDR56OV7DXYRB","short_pith_number":"pith:XWQRXIRX","schema_version":"1.0","canonical_sha256":"bda11ba23775a9944071ef9d5f8ef8886c80de0bef8553131c6dec5b4d344429","source":{"kind":"arxiv","id":"2112.04491","version":4},"attestation_state":"computed","paper":{"title":"Improving Image Restoration by Revisiting Global Information Aggregation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["eess.IV"],"primary_cat":"cs.CV","authors_text":"Chengpeng Chen, Liangyu Chen, Xiaojie Chu, Xin Lu","submitted_at":"2021-12-08T12:52:14Z","abstract_excerpt":"Global operations, such as global average pooling, are widely used in top-performance image restorers. They aggregate global information from input features along entire spatial dimensions but behave differently during training and inference in image restoration tasks: they are based on different regions, namely the cropped patches (from images) and the full-resolution images. This paper revisits global information aggregation and finds that the image-based features during inference have a different distribution than the patch-based features during training. This train-test inconsistency negat"},"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":"2112.04491","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2021-12-08T12:52:14Z","cross_cats_sorted":["eess.IV"],"title_canon_sha256":"387050415a194aa67b1088e9d34ebc92afe36ace0306ab560b8137ae76e6c15e","abstract_canon_sha256":"1a89def2e730e90205cf649c0025575622964b19ebfeadb8d338c6f71fbd9f97"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:46:34.056011Z","signature_b64":"90xHx2ksFl14VzSUej1Oyw8czH9NZZ+of8Cs9vUwjW2yUHzhXCmBKIybcjr5ICb4O4xTHBOnksVhVgiuzMusCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bda11ba23775a9944071ef9d5f8ef8886c80de0bef8553131c6dec5b4d344429","last_reissued_at":"2026-07-05T04:46:34.055676Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:46:34.055676Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Image Restoration by Revisiting Global Information Aggregation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["eess.IV"],"primary_cat":"cs.CV","authors_text":"Chengpeng Chen, Liangyu Chen, Xiaojie Chu, Xin Lu","submitted_at":"2021-12-08T12:52:14Z","abstract_excerpt":"Global operations, such as global average pooling, are widely used in top-performance image restorers. They aggregate global information from input features along entire spatial dimensions but behave differently during training and inference in image restoration tasks: they are based on different regions, namely the cropped patches (from images) and the full-resolution images. This paper revisits global information aggregation and finds that the image-based features during inference have a different distribution than the patch-based features during training. This train-test inconsistency negat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.04491","kind":"arxiv","version":4},"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/2112.04491/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":"2112.04491","created_at":"2026-07-05T04:46:34.055730+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.04491v4","created_at":"2026-07-05T04:46:34.055730+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.04491","created_at":"2026-07-05T04:46:34.055730+00:00"},{"alias_kind":"pith_short_12","alias_value":"XWQRXIRXOWUZ","created_at":"2026-07-05T04:46:34.055730+00:00"},{"alias_kind":"pith_short_16","alias_value":"XWQRXIRXOWUZIQDR","created_at":"2026-07-05T04:46:34.055730+00:00"},{"alias_kind":"pith_short_8","alias_value":"XWQRXIRX","created_at":"2026-07-05T04:46:34.055730+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.18403","citing_title":"Efficient Transformer for High Resolution Image Motion Deblurring","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB","json":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB.json","graph_json":"https://pith.science/api/pith-number/XWQRXIRXOWUZIQDR56OV7DXYRB/graph.json","events_json":"https://pith.science/api/pith-number/XWQRXIRXOWUZIQDR56OV7DXYRB/events.json","paper":"https://pith.science/paper/XWQRXIRX"},"agent_actions":{"view_html":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB","download_json":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB.json","view_paper":"https://pith.science/paper/XWQRXIRX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.04491&json=true","fetch_graph":"https://pith.science/api/pith-number/XWQRXIRXOWUZIQDR56OV7DXYRB/graph.json","fetch_events":"https://pith.science/api/pith-number/XWQRXIRXOWUZIQDR56OV7DXYRB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB/action/storage_attestation","attest_author":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB/action/author_attestation","sign_citation":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB/action/citation_signature","submit_replication":"https://pith.science/pith/XWQRXIRXOWUZIQDR56OV7DXYRB/action/replication_record"}},"created_at":"2026-07-05T04:46:34.055730+00:00","updated_at":"2026-07-05T04:46:34.055730+00:00"}