{"corpus_id":40094999,"paper_sha":"b8e1914c78c0b616e7d081759b1343cbfead42ad","doi":null,"arxiv_id":"1711.05225","pmid":null,"pmcid":null,"mag_id":2770241596,"dblp_id":"journals/corr/abs-1711-05225","acl_id":null,"title":"CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning","year":2017,"publication_date":"2017-11-14","venue":"arXiv.org","journal":{"name":"ArXiv","pages":null,"volume":"abs/1711.05225"},"journal_issn":null,"journal_title":null,"publication_types":["JournalArticle"],"pubmed_pub_types":null,"s2_fields_of_study":["Medicine","Computer Science","Mathematics"],"reference_count":30,"citation_count":3129,"influential_citation_count":320,"is_open_access":false,"arxiv_categories":["cs.CV","cs.LG","stat.ML"],"arxiv_license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","arxiv_journal_ref":null,"mesh_headings":null,"chemicals":null,"comments_corrections":null,"source_flags":1,"s2_open_access_pdf_url":null,"s2_open_access_landing_url":null,"s2_open_access_license":null,"s2_open_access_status":null,"pmc_open_access_pdf_url":null,"pmc_open_access_landing_url":null,"pmc_open_access_license":null,"pmc_open_access_status":null,"unpaywall_open_access_pdf_url":null,"unpaywall_open_access_landing_url":null,"unpaywall_open_access_license":null,"unpaywall_open_access_status":null,"abstract":"We develop an algorithm that can detect pneumonia from chest X-rays at a level exceeding practicing radiologists. Our algorithm, CheXNet, is a 121-layer convolutional neural network trained on ChestX-ray14, currently the largest publicly available chest X-ray dataset, containing over 100,000 frontal-view X-ray images with 14 diseases. Four practicing academic radiologists annotate a test set, on which we compare the performance of CheXNet to that of radiologists. We find that CheXNet exceeds average radiologist performance on the F1 metric. We extend CheXNet to detect all 14 diseases in ChestX-ray14 and achieve state of the art results on all 14 diseases.","claims":[{"public_id":"cl_50744f40a0e52298dc1b817ce0599d2b","status":"active","text":"CheXNet detects pneumonia from chest X-rays at a level exceeding practicing radiologists.","confidence":0.98,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_50744f40a0e52298dc1b817ce0599d2b"},{"public_id":"cl_a1cd4dfea4799f64ff59ff78b2539f10","status":"active","text":"CheXNet exceeds average radiologist performance on the F1 metric on the radiologist-annotated test set.","confidence":0.97,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_a1cd4dfea4799f64ff59ff78b2539f10"},{"public_id":"cl_7a72595d430737c963adef9e98770e0b","status":"active","text":"CheXNet is a 121-layer convolutional neural network trained on ChestX-ray14.","confidence":0.98,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_7a72595d430737c963adef9e98770e0b"},{"public_id":"cl_3d1ef2e7791cfe914bc70611bef491c5","status":"active","text":"The extended CheXNet model achieves state-of-the-art results on all 14 diseases in ChestX-ray14.","confidence":0.96,"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/claims/cl_3d1ef2e7791cfe914bc70611bef491c5"}],"concepts":[{"public_id":"co_4217d6ca37c01dd959b9e322a4efe190","status":"active","name":"practicing academic radiologists","description":"Academic radiologists who contributed annotations for the test set and served as a comparison group.","types":["expert group"],"aliases":["academic radiologists"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_4217d6ca37c01dd959b9e322a4efe190"},{"public_id":"co_4bc18893f3dfafea17a8bd5a38aeecf5","status":"active","name":"chest X-rays","description":"Frontal-view chest radiographic images used as input for diagnosis.","types":["medical imaging modality"],"aliases":["chest radiographs","CXR"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_4bc18893f3dfafea17a8bd5a38aeecf5"},{"public_id":"co_518a91d77028f0dfde727aa992088ff9","status":"active","name":"ChestX-ray14","description":"A publicly available chest X-ray dataset containing more than 100,000 frontal-view images labeled for 14 diseases.","types":["dataset"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_518a91d77028f0dfde727aa992088ff9"},{"public_id":"co_6060ebcca05a672fb6d1a4fa03688025","status":"active","name":"practicing radiologists","description":"Clinicians who interpret medical images in routine practice.","types":["expert group"],"aliases":["radiologists"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_6060ebcca05a672fb6d1a4fa03688025"},{"public_id":"co_6776f08d9d6a8ada7707d81966fa0354","status":"active","name":"radiologist-annotated test set","description":"A held-out test set labeled by four practicing academic radiologists for performance comparison.","types":["evaluation dataset"],"aliases":["test set"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_6776f08d9d6a8ada7707d81966fa0354"},{"public_id":"co_6b5c6c4eab153407be2da797fecd3868","status":"active","name":"14 diseases","description":"The set of thoracic disease labels present in ChestX-ray14 and used for multi-label detection.","types":["label set"],"aliases":["all 14 diseases"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_6b5c6c4eab153407be2da797fecd3868"},{"public_id":"co_a6813eabe5703c5e21c93c21b829e868","status":"active","name":"121-layer convolutional neural network","description":"A deep convolutional neural network architecture with 121 layers.","types":["model architecture"],"aliases":["121-layer CNN"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_a6813eabe5703c5e21c93c21b829e868"},{"public_id":"co_b015156adba675d58f5aed919133c0b5","status":"active","name":"CheXNet","description":"A deep learning system for detecting pneumonia and other thoracic diseases from chest X-ray images.","types":["model","method"],"aliases":[],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_b015156adba675d58f5aed919133c0b5"},{"public_id":"co_eb6a15debf9299dee7198c5f4fa69cbc","status":"active","name":"pneumonia detection","description":"The task of identifying pneumonia from medical imaging, especially chest X-rays.","types":["task"],"aliases":["detect pneumonia"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_eb6a15debf9299dee7198c5f4fa69cbc"},{"public_id":"co_ec0e0a3beaab0322bdfb08c5e1571d6f","status":"active","name":"F1 metric","description":"A performance measure combining precision and recall.","types":["evaluation metric"],"aliases":["F1 score"],"contributors":[{"id":1,"public_id":"12632b8b5f","public_label":"Anonymous (12632b8b5f)","roles":["extraction"],"url":"https://sah.borca.ai/u/12632b8b5f"}],"url":"https://sah.borca.ai/concepts/co_ec0e0a3beaab0322bdfb08c5e1571d6f"}],"external_ids":{"DOI":null,"ArXiv":"1711.05225","PubMed":null,"PubMedCentral":null,"MAG":2770241596,"DBLP":"journals/corr/abs-1711-05225","ACL":null},"open_access":{"is_open_access":true,"pdf_url":"https://arxiv.org/pdf/1711.05225","landing_url":"https://arxiv.org/abs/1711.05225","source":"arxiv","pdf_url_source":"derived_arxiv","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","reason":null},"reference_availability":{"status":"available","references_indexed":true,"full_text_available":true,"full_text_source":"arxiv","count_basis":"semantic_scholar_metadata","extraction_status":"not_applicable","reason":null},"source":{"provider":"episteme2","base_corpus":"semantic_scholar_dump","freshness_mode":"unknown","basis":["semantic_scholar_metadata","postgres_metadata"],"limits":["paper metadata is based on indexed upstream scholarly datasets","claims and concepts are available only for extracted papers","absence of claims or concepts means no extracted graph data is available in this response"],"status":"available","degraded":false,"degraded_reasons":[],"diagnostics":{"status":"available","degraded":false,"degraded_reasons":[],"metadata_status":"available","graph_status":"available","abstract_status":"available"},"source_flags":1},"paper_id":636628,"paper_uid":"36409b3f-2dd4-4af8-bccf-c1dc0ac66298","canonical_identity":{"paper_id":636628,"paper_uid":"36409b3f-2dd4-4af8-bccf-c1dc0ac66298","identity_status":"available","lookup_basis":"semantic_scholar_external_id","compatibility_path":"corpus_id"},"url":"https://sah.borca.ai/papers/40094999"}