{"paper":{"title":"MedP-CLIP: Medical CLIP with Region-Aware Prompt Integration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MedP-CLIP adds a feature-level region prompt mechanism to CLIP for medical images.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"He Yao, Hongchun Lu, Jiahui Peng, Jingwen Li, Jin Ye, Junlong Cheng, Lincheng Jiang, Min Zhu, Sibo Ju, Xue Li, Yanzhou Su, Yujie Lu","submitted_at":"2026-04-13T08:53:36Z","abstract_excerpt":"Contrastive Language-Image Pre-training (CLIP) has demonstrated outstanding performance in global image understanding and zero-shot transfer through large-scale text-image alignment. However, the core of medical image analysis often lies in the fine-grained understanding of specific anatomical structures or lesion regions. Therefore, precisely comprehending region-of-interest (RoI) information provided by medical professionals or perception models becomes crucial. To address this need, we propose MedP-CLIP, a region-aware medical vision-language model (VLM). MedP-CLIP innovatively integrates m"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MedP-CLIP significantly outperforms baseline methods in various medical tasks, including zero-shot recognition, interactive segmentation, and empowering multimodal large language models. This model provides a scalable, plug-and-play visual backbone for medical AI, combining holistic image understanding with precise regional analysis.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The feature-level region prompt integration mechanism enables it to flexibly respond to various prompt forms (e.g., points, bounding boxes, masks) while maintaining global contextual awareness when focusing on local regions.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MedP-CLIP is a medical CLIP variant with feature-level region prompt integration, pre-trained on 6.4 million images and 97.3 million region annotations, that outperforms baselines in zero-shot recognition and interactive segmentation.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MedP-CLIP adds a feature-level region prompt mechanism to CLIP for medical images.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"11e7bc0b627b878b90658b9776078cd8433840f9c0d7f524d724f069859b717b"},"source":{"id":"2604.11197","kind":"arxiv","version":2},"verdict":{"id":"b92f995b-b78c-48a6-b20e-7bd1ffa7798d","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T15:50:56.118485Z","strongest_claim":"MedP-CLIP significantly outperforms baseline methods in various medical tasks, including zero-shot recognition, interactive segmentation, and empowering multimodal large language models. This model provides a scalable, plug-and-play visual backbone for medical AI, combining holistic image understanding with precise regional analysis.","one_line_summary":"MedP-CLIP is a medical CLIP variant with feature-level region prompt integration, pre-trained on 6.4 million images and 97.3 million region annotations, that outperforms baselines in zero-shot recognition and interactive segmentation.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The feature-level region prompt integration mechanism enables it to flexibly respond to various prompt forms (e.g., points, bounding boxes, masks) while maintaining global contextual awareness when focusing on local regions.","pith_extraction_headline":"MedP-CLIP adds a feature-level region prompt mechanism to CLIP for medical images."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.11197/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"}