{"paper":{"title":"UHR-Net: An Uncertainty-Aware Hypergraph Refinement Network for Medical Image Segmentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"UHR-Net refines lesion boundaries by splitting hyperedge prototypes according to an entropy uncertainty map.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jinghao Shi, Kun Sun, Shuokun Cheng","submitted_at":"2026-04-30T16:38:51Z","abstract_excerpt":"Accurate lesion segmentation is crucial for clinical diagnosis and treatment planning. However, lesions often resemble surrounding tissues and exhibit ill-defined boundaries, leading to unstable predictions in boundary/transition regions. Moreover, small-lesion cues can be diluted by multi-scale feature extraction, causing under- or over-segmentation. To address these challenges, we propose an Uncertainty-Aware Hypergraph Refinement Network (UHR-Net). First, we introduce an Uncertainty-Oriented Instance Contrastive (UO-IC) pretraining strategy that couples geometry-aware copy-paste augmentatio"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"By splitting hyperedge prototypes into foreground and background groups, UGHR decouples higher-order interactions and improves refinement in ambiguous regions. Experiments on five public benchmarks demonstrate consistent gains over strong baselines.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That an entropy-based uncertainty map derived from a coarse probability map reliably identifies ambiguous regions and that guiding hypergraph refinement with it will not propagate errors from the initial coarse prediction.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"UHR-Net improves medical lesion segmentation accuracy by using uncertainty-guided hypergraph refinement and instance contrastive pretraining to better handle ambiguous boundaries and small lesions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"UHR-Net refines lesion boundaries by splitting hyperedge prototypes according to an entropy uncertainty map.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"debc249ab1562c4b8aca108a75d644993ba04cf51c5ad394f60b4d41ef7efa1e"},"source":{"id":"2604.28095","kind":"arxiv","version":2},"verdict":{"id":"bbecb2a0-bbe9-476f-80c4-73e685e1a118","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-07T04:53:08.470279Z","strongest_claim":"By splitting hyperedge prototypes into foreground and background groups, UGHR decouples higher-order interactions and improves refinement in ambiguous regions. Experiments on five public benchmarks demonstrate consistent gains over strong baselines.","one_line_summary":"UHR-Net improves medical lesion segmentation accuracy by using uncertainty-guided hypergraph refinement and instance contrastive pretraining to better handle ambiguous boundaries and small lesions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That an entropy-based uncertainty map derived from a coarse probability map reliably identifies ambiguous regions and that guiding hypergraph refinement with it will not propagate errors from the initial coarse prediction.","pith_extraction_headline":"UHR-Net refines lesion boundaries by splitting hyperedge prototypes according to an entropy uncertainty map."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.28095/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T20:41:05.075426Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T18:37:03.676047Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"a4d02210662a5bad244e8fa350cef4ac53c15dfddfcfccc7c27ad7308cabf9d3"},"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"}