{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CAPME2KJDFSMKE7VRWNIHUC6LX","short_pith_number":"pith:CAPME2KJ","schema_version":"1.0","canonical_sha256":"101ec269491964c513f58d9a83d05e5ddbae08b03079464f5a86567981c4b606","source":{"kind":"arxiv","id":"2604.19191","version":2},"attestation_state":"computed","paper":{"title":"Towards Modality-Agnostic Medical Image Anomaly Detection: A Training-Free Manifold Refinement Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Mean Shift Density Enhancement refines latent features from pretrained backbones to sharpen one-class anomaly scoring in medical images.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Gouri Lakshmi S, Pritam Kar, Saptarshi Bej","submitted_at":"2026-04-21T07:58:44Z","abstract_excerpt":"Deploying AI-based anomaly detection across diverse clinical imaging settings remains challenging because most existing methods rely on modality-specific architectures, anatomical priors, or extensive retraining, limiting their use as general-purpose screening tools. One-class classification (OCC) offers a label-efficient alternative by training exclusively on normal data, but conventional two-stage pipelines fit a density estimator directly on raw pretrained embeddings, leaving substantial discriminative structure in the latent space unexploited. We introduce a training-free, modality-agnosti"},"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":"2604.19191","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-04-21T07:58:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"2f350d25ff3cb5168fd6facd99724f111f661f8c30f422cd5f2cdacf7039baf8","abstract_canon_sha256":"9d0eae0cb8656a187a83f681977a39ffdfff081f23d0f37b632aa768aa5381da"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:16:30.055327Z","signature_b64":"d5XiiFPTs7RCUcHdozkMV0xDvygm/tDX0wSKHTOgyZVP67c1H9NN+9taj3TdaFx897maEBZ6Emfu2nAU3LpnCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"101ec269491964c513f58d9a83d05e5ddbae08b03079464f5a86567981c4b606","last_reissued_at":"2026-06-30T01:16:30.054884Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:16:30.054884Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Modality-Agnostic Medical Image Anomaly Detection: A Training-Free Manifold Refinement Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Mean Shift Density Enhancement refines latent features from pretrained backbones to sharpen one-class anomaly scoring in medical images.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Gouri Lakshmi S, Pritam Kar, Saptarshi Bej","submitted_at":"2026-04-21T07:58:44Z","abstract_excerpt":"Deploying AI-based anomaly detection across diverse clinical imaging settings remains challenging because most existing methods rely on modality-specific architectures, anatomical priors, or extensive retraining, limiting their use as general-purpose screening tools. One-class classification (OCC) offers a label-efficient alternative by training exclusively on normal data, but conventional two-stage pipelines fit a density estimator directly on raw pretrained embeddings, leaving substantial discriminative structure in the latent space unexploited. We introduce a training-free, modality-agnosti"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MSDE achieves the highest AUC on four datasets and the highest Average Precision on five datasets, including near-perfect performance on brain tumor detection (0.981 AUC/AP).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the iterative manifold-shifting performed by Mean Shift Density Enhancement reliably moves normal samples toward higher-likelihood regions in the latent space of arbitrary pretrained backbones, thereby sharpening the subsequent Gaussian density estimate for anomaly scoring.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A new Mean Shift Density Enhancement procedure applied to self-supervised embeddings yields state-of-the-art anomaly detection AUC and average precision on seven medical imaging datasets.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Mean Shift Density Enhancement refines latent features from pretrained backbones to sharpen one-class anomaly scoring in medical images.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d2883a1417a4b6b7c6552da9436432e6ad0e72bc3b914db1cb65b82d3a046184"},"source":{"id":"2604.19191","kind":"arxiv","version":2},"verdict":{"id":"c3ac9aa1-b4de-4580-afa5-3c6fbfebd6bd","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T02:46:07.745809Z","strongest_claim":"MSDE achieves the highest AUC on four datasets and the highest Average Precision on five datasets, including near-perfect performance on brain tumor detection (0.981 AUC/AP).","one_line_summary":"A new Mean Shift Density Enhancement procedure applied to self-supervised embeddings yields state-of-the-art anomaly detection AUC and average precision on seven medical imaging datasets.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the iterative manifold-shifting performed by Mean Shift Density Enhancement reliably moves normal samples toward higher-likelihood regions in the latent space of arbitrary pretrained backbones, thereby sharpening the subsequent Gaussian density estimate for anomaly scoring.","pith_extraction_headline":"Mean Shift Density Enhancement refines latent features from pretrained backbones to sharpen one-class anomaly scoring in medical images."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.19191/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-20T03:14:49.843223Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"5cb34724f26ca1f229b6b2aa64dcb9e9ed700b3d718b37b617416e563237ecbe"},"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":"2604.19191","created_at":"2026-06-30T01:16:30.054946+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.19191v2","created_at":"2026-06-30T01:16:30.054946+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.19191","created_at":"2026-06-30T01:16:30.054946+00:00"},{"alias_kind":"pith_short_12","alias_value":"CAPME2KJDFSM","created_at":"2026-06-30T01:16:30.054946+00:00"},{"alias_kind":"pith_short_16","alias_value":"CAPME2KJDFSMKE7V","created_at":"2026-06-30T01:16:30.054946+00:00"},{"alias_kind":"pith_short_8","alias_value":"CAPME2KJ","created_at":"2026-06-30T01:16:30.054946+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/CAPME2KJDFSMKE7VRWNIHUC6LX","json":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX.json","graph_json":"https://pith.science/api/pith-number/CAPME2KJDFSMKE7VRWNIHUC6LX/graph.json","events_json":"https://pith.science/api/pith-number/CAPME2KJDFSMKE7VRWNIHUC6LX/events.json","paper":"https://pith.science/paper/CAPME2KJ"},"agent_actions":{"view_html":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX","download_json":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX.json","view_paper":"https://pith.science/paper/CAPME2KJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.19191&json=true","fetch_graph":"https://pith.science/api/pith-number/CAPME2KJDFSMKE7VRWNIHUC6LX/graph.json","fetch_events":"https://pith.science/api/pith-number/CAPME2KJDFSMKE7VRWNIHUC6LX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX/action/storage_attestation","attest_author":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX/action/author_attestation","sign_citation":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX/action/citation_signature","submit_replication":"https://pith.science/pith/CAPME2KJDFSMKE7VRWNIHUC6LX/action/replication_record"}},"created_at":"2026-06-30T01:16:30.054946+00:00","updated_at":"2026-06-30T01:16:30.054946+00:00"}