{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DU3Q5YQDKSWON3Z4A3AOWAT55G","short_pith_number":"pith:DU3Q5YQD","schema_version":"1.0","canonical_sha256":"1d370ee20354ace6ef3c06c0eb027de99f13203bbd7e5537040c936402780a9b","source":{"kind":"arxiv","id":"2309.09668","version":2},"attestation_state":"computed","paper":{"title":"DFormer: Rethinking RGBD Representation Learning for Semantic Segmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bowen Yin, Li Liu, Ming-Ming Cheng, Qibin Hou, Xuying Zhang, Zhongyu Li","submitted_at":"2023-09-18T11:09:11Z","abstract_excerpt":"We present DFormer, a novel RGB-D pretraining framework to learn transferable representations for RGB-D segmentation tasks. DFormer has two new key innovations: 1) Unlike previous works that encode RGB-D information with RGB pretrained backbone, we pretrain the backbone using image-depth pairs from ImageNet-1K, and hence the DFormer is endowed with the capacity to encode RGB-D representations; 2) DFormer comprises a sequence of RGB-D blocks, which are tailored for encoding both RGB and depth information through a novel building block design. DFormer avoids the mismatched encoding of the 3D geo"},"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":"2309.09668","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-09-18T11:09:11Z","cross_cats_sorted":[],"title_canon_sha256":"9910a54c32ec66359678d2302c5616333614d0fdea0938021bd8de88e26ca5fd","abstract_canon_sha256":"833fa065298051ab97ce28bee457507cbd56097f4115c922770013987eba6702"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:42:20.024049Z","signature_b64":"CigGRN8d0ShXKk3Fuvnt3HypwLTgL+c0qugdM3ZErN49w804wZRqX2Jx31NKL0/6L73fU/LFTmq1vMPITof/BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d370ee20354ace6ef3c06c0eb027de99f13203bbd7e5537040c936402780a9b","last_reissued_at":"2026-07-05T07:42:20.023565Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:42:20.023565Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DFormer: Rethinking RGBD Representation Learning for Semantic Segmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bowen Yin, Li Liu, Ming-Ming Cheng, Qibin Hou, Xuying Zhang, Zhongyu Li","submitted_at":"2023-09-18T11:09:11Z","abstract_excerpt":"We present DFormer, a novel RGB-D pretraining framework to learn transferable representations for RGB-D segmentation tasks. DFormer has two new key innovations: 1) Unlike previous works that encode RGB-D information with RGB pretrained backbone, we pretrain the backbone using image-depth pairs from ImageNet-1K, and hence the DFormer is endowed with the capacity to encode RGB-D representations; 2) DFormer comprises a sequence of RGB-D blocks, which are tailored for encoding both RGB and depth information through a novel building block design. DFormer avoids the mismatched encoding of the 3D geo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2309.09668","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/2309.09668/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":"2309.09668","created_at":"2026-07-05T07:42:20.023621+00:00"},{"alias_kind":"arxiv_version","alias_value":"2309.09668v2","created_at":"2026-07-05T07:42:20.023621+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2309.09668","created_at":"2026-07-05T07:42:20.023621+00:00"},{"alias_kind":"pith_short_12","alias_value":"DU3Q5YQDKSWO","created_at":"2026-07-05T07:42:20.023621+00:00"},{"alias_kind":"pith_short_16","alias_value":"DU3Q5YQDKSWON3Z4","created_at":"2026-07-05T07:42:20.023621+00:00"},{"alias_kind":"pith_short_8","alias_value":"DU3Q5YQD","created_at":"2026-07-05T07:42:20.023621+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.13206","citing_title":"Visual Place Recognition in Forests with Depth-Aware Distillation","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2407.11906","citing_title":"SegSTRONG-C: Segmenting Surgical Tools Robustly On Non-adversarial Generated Corruptions -- An EndoVis'24 Challenge","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02948","citing_title":"CrossWeaver: Cross-modal Weaving for Arbitrary-Modality Semantic Segmentation","ref_index":37,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G","json":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G.json","graph_json":"https://pith.science/api/pith-number/DU3Q5YQDKSWON3Z4A3AOWAT55G/graph.json","events_json":"https://pith.science/api/pith-number/DU3Q5YQDKSWON3Z4A3AOWAT55G/events.json","paper":"https://pith.science/paper/DU3Q5YQD"},"agent_actions":{"view_html":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G","download_json":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G.json","view_paper":"https://pith.science/paper/DU3Q5YQD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2309.09668&json=true","fetch_graph":"https://pith.science/api/pith-number/DU3Q5YQDKSWON3Z4A3AOWAT55G/graph.json","fetch_events":"https://pith.science/api/pith-number/DU3Q5YQDKSWON3Z4A3AOWAT55G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G/action/storage_attestation","attest_author":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G/action/author_attestation","sign_citation":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G/action/citation_signature","submit_replication":"https://pith.science/pith/DU3Q5YQDKSWON3Z4A3AOWAT55G/action/replication_record"}},"created_at":"2026-07-05T07:42:20.023621+00:00","updated_at":"2026-07-05T07:42:20.023621+00:00"}