{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TVCLVDSLUAFINYIQ3AQETNC3YR","short_pith_number":"pith:TVCLVDSL","schema_version":"1.0","canonical_sha256":"9d44ba8e4ba00a86e110d82049b45bc46a6e97315478183141f90004ff2af2e4","source":{"kind":"arxiv","id":"2605.14403","version":1},"attestation_state":"computed","paper":{"title":"DermAgent: A Self-Reflective Agentic System for Dermatological Image Analysis with Multi-Tool Reasoning and Traceable Decision-Making","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"DermAgent anchors each skin image prediction in retrieved cases and guidelines then self-corrects via critic gates to raise diagnostic accuracy above standard models.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Feilong Tang, Lie Ju, Ming Hu, Siyuan Yan, Wei Feng, Xieji Li, Yize Liu, Zongyuan Ge","submitted_at":"2026-05-14T05:41:11Z","abstract_excerpt":"Dermatological diagnosis requires integrating fine-grained visual perception with expert clinical knowledge. Although Multimodal Large Language Models (MLLMs) facilitate interactive medical image analysis, their application in dermatology is hindered by insufficient domain-specific grounding and hallucinations. To address these issues, we propose DermAgent, a collaborative multi-tool agent that orchestrates seven specialized vision and language modules within a Plan-Execute-Reflect framework. DermAgent delivers stepwise, traceable diagnostic reasoning through three core components. First, it e"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.14403","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-05-14T05:41:11Z","cross_cats_sorted":[],"title_canon_sha256":"d42a931dedcc7a55c3a6f1c75d7c51e5f1b2fbcc27c72312240bae7fded11678","abstract_canon_sha256":"ecbce4e7e338026f359f1072060c5209f7c4de3e364bbe662ecb7dbfffaf49cf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:07.460596Z","signature_b64":"xiIRWvuLMXgZhn+NaYNmN9WCMguxUSbLYYuZuf7MypxTdcpJgBs5NXAcG8cM7ANt1DyUOJ0MtyjfSgTQNE1bDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9d44ba8e4ba00a86e110d82049b45bc46a6e97315478183141f90004ff2af2e4","last_reissued_at":"2026-05-17T23:39:07.459967Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:07.459967Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DermAgent: A Self-Reflective Agentic System for Dermatological Image Analysis with Multi-Tool Reasoning and Traceable Decision-Making","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"DermAgent anchors each skin image prediction in retrieved cases and guidelines then self-corrects via critic gates to raise diagnostic accuracy above standard models.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Feilong Tang, Lie Ju, Ming Hu, Siyuan Yan, Wei Feng, Xieji Li, Yize Liu, Zongyuan Ge","submitted_at":"2026-05-14T05:41:11Z","abstract_excerpt":"Dermatological diagnosis requires integrating fine-grained visual perception with expert clinical knowledge. Although Multimodal Large Language Models (MLLMs) facilitate interactive medical image analysis, their application in dermatology is hindered by insufficient domain-specific grounding and hallucinations. To address these issues, we propose DermAgent, a collaborative multi-tool agent that orchestrates seven specialized vision and language modules within a Plan-Execute-Reflect framework. DermAgent delivers stepwise, traceable diagnostic reasoning through three core components. First, it e"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"DermAgent consistently outperforms state-of-the-art MLLMs and medical agent baselines across zero-shot fine-grained disease diagnosis, concept annotation, and clinical captioning tasks, exceeding GPT-4o by 17.6% in skin disease diagnostic accuracy and 3.15% in captioning ROUGE-L.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The dual-modality retrieval from 413,210 diagnosed cases and 3,199 guideline chunks provides unbiased, comprehensive anchoring for every prediction, and the critic module's confidence-coverage-conflict gates reliably detect and correct hallucinations without introducing new errors or over-correction.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DermAgent orchestrates seven vision-language tools in a Plan-Execute-Reflect loop with dual-modality retrieval from 413k cases and a critic module to outperform GPT-4o by 17.6% in zero-shot dermatological diagnosis accuracy.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DermAgent anchors each skin image prediction in retrieved cases and guidelines then self-corrects via critic gates to raise diagnostic accuracy above standard models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"8acccfbf050f0deb092646722b2755d63fd1153ca71df69473a1cfa746866a10"},"source":{"id":"2605.14403","kind":"arxiv","version":1},"verdict":{"id":"393ca0aa-9def-41c7-b7ac-1b98ef12239e","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T02:52:42.920869Z","strongest_claim":"DermAgent consistently outperforms state-of-the-art MLLMs and medical agent baselines across zero-shot fine-grained disease diagnosis, concept annotation, and clinical captioning tasks, exceeding GPT-4o by 17.6% in skin disease diagnostic accuracy and 3.15% in captioning ROUGE-L.","one_line_summary":"DermAgent orchestrates seven vision-language tools in a Plan-Execute-Reflect loop with dual-modality retrieval from 413k cases and a critic module to outperform GPT-4o by 17.6% in zero-shot dermatological diagnosis accuracy.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The dual-modality retrieval from 413,210 diagnosed cases and 3,199 guideline chunks provides unbiased, comprehensive anchoring for every prediction, and the critic module's confidence-coverage-conflict gates reliably detect and correct hallucinations without introducing new errors or over-correction.","pith_extraction_headline":"DermAgent anchors each skin image prediction in retrieved cases and guidelines then self-corrects via critic gates to raise diagnostic accuracy above standard models."},"references":{"count":33,"sample":[{"doi":"","year":null,"title":"https://dermnetnz.org/","work_id":"c42c6104-02f6-4e95-88db-a5b846a88c91","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"https://www.mayoclinic.org/diseases-conditions","work_id":"930fa1dc-52ff-40c2-9863-f5d00b782a56","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.48550/arxiv.2511.21631","year":2025,"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","ref_index":3,"cited_arxiv_id":"2511.21631","is_internal_anchor":true},{"doi":"10.48550/arxiv.2406.19280","year":2024,"title":"Huatuogpt-vision, towards injecting medical visual knowledge into multimodal llms at scale","work_id":"dd32b8a1-4ad4-4155-b031-c317b565c6e7","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.48550/arxiv.2302.00785","year":2023,"title":"https://doi.org/10.48550/arXiv.2302.00785","work_id":"307faa4a-126f-4df4-a7c4-0e19d8db60d2","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":33,"snapshot_sha256":"f4c7ac85d904bc957c5366881b68a6a142f0de73c98b965bc82849200b6a9da8","internal_anchors":4},"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":"2605.14403","created_at":"2026-05-17T23:39:07.460075+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.14403v1","created_at":"2026-05-17T23:39:07.460075+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14403","created_at":"2026-05-17T23:39:07.460075+00:00"},{"alias_kind":"pith_short_12","alias_value":"TVCLVDSLUAFI","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"TVCLVDSLUAFINYIQ","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"TVCLVDSL","created_at":"2026-05-18T12:33:37.589309+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/TVCLVDSLUAFINYIQ3AQETNC3YR","json":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR.json","graph_json":"https://pith.science/api/pith-number/TVCLVDSLUAFINYIQ3AQETNC3YR/graph.json","events_json":"https://pith.science/api/pith-number/TVCLVDSLUAFINYIQ3AQETNC3YR/events.json","paper":"https://pith.science/paper/TVCLVDSL"},"agent_actions":{"view_html":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR","download_json":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR.json","view_paper":"https://pith.science/paper/TVCLVDSL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.14403&json=true","fetch_graph":"https://pith.science/api/pith-number/TVCLVDSLUAFINYIQ3AQETNC3YR/graph.json","fetch_events":"https://pith.science/api/pith-number/TVCLVDSLUAFINYIQ3AQETNC3YR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR/action/storage_attestation","attest_author":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR/action/author_attestation","sign_citation":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR/action/citation_signature","submit_replication":"https://pith.science/pith/TVCLVDSLUAFINYIQ3AQETNC3YR/action/replication_record"}},"created_at":"2026-05-17T23:39:07.460075+00:00","updated_at":"2026-05-17T23:39:07.460075+00:00"}