@misc{rajpurkar2017dataset,
  abstract = {We introduce MURA, a large dataset of musculoskeletal radiographs containing
40,895 images from 14,982 studies, where each study is manually labeled by
radiologists as either normal or abnormal. On this dataset, we train a
169-layer densely connected convolutional network to detect and localize
abnormalities. To evaluate our model robustly and to get an estimate of
radiologist performance, we collect additional labels from board-certified
Stanford radiologists on the test set, consisting of 209 musculoskeletal
studies. We compared our model and radiologists on the Cohen's kappa statistic,
which expresses the agreement of our model and of each radiologist with the
gold standard, defined as the majority vote of a disjoint group of
radiologists. We find that our model achieves performance comparable to that of
radiologists. Model performance is higher than the best radiologist performance
in detecting abnormalities on finger studies and equivalent on wrist studies.
However, model performance is lower than best radiologist performance in
detecting abnormalities on elbow, forearm, hand, humerus, and shoulder studies,
indicating that the task is a good challenge for future research. To encourage
advances, we have made our dataset freely available at
https://stanfordmlgroup.github.io/projects/mura},
  added-at = {2018-01-19T10:15:17.000+0100},
  author = {Rajpurkar, Pranav and Irvin, Jeremy and Bagul, Aarti and Ding, Daisy and Duan, Tony and Mehta, Hershel and Yang, Brandon and Zhu, Kaylie and Laird, Dillon and Ball, Robyn L. and Langlotz, Curtis and Shpanskaya, Katie and Lungren, Matthew P. and Ng, Andrew},
  biburl = {https://www.bibsonomy.org/bibtex/2c8ff2f68a0e57716153eefdda755bfd8/alexattia},
  description = {MURA Dataset: Towards Radiologist-Level Abnormality Detection in
  Musculoskeletal Radiographs},
  interhash = {0fbf86d39e14c19e59ea2d81d3feda51},
  intrahash = {c8ff2f68a0e57716153eefdda755bfd8},
  keywords = {azmed deep-learning recvis},
  note = {cite arxiv:1712.06957},
  timestamp = {2018-01-19T10:15:17.000+0100},
  title = {MURA Dataset: Towards Radiologist-Level Abnormality Detection in
  Musculoskeletal Radiographs},
  url = {http://arxiv.org/abs/1712.06957},
  year = 2017
}

{"_id":"fMkSB4Cj5BmXHwp8x","bibbaseid":"rajpurkar-irvin-bagul-ding-duan-mehta-yang-zhu-etal-muradatasettowardsradiologistlevelabnormalitydetectioninmusculoskeletalradiographs-2017","author_short":["Rajpurkar, P.","Irvin, J.","Bagul, A.","Ding, D.","Duan, T.","Mehta, H.","Yang, B.","Zhu, K.","Laird, D.","Ball, R. L.","Langlotz, C.","Shpanskaya, K.","Lungren, M. P.","Ng, A."],"bibdata":{"bibtype":"misc","type":"misc","abstract":"We introduce MURA, a large dataset of musculoskeletal radiographs containing 40,895 images from 14,982 studies, where each study is manually labeled by radiologists as either normal or abnormal. On this dataset, we train a 169-layer densely connected convolutional network to detect and localize abnormalities. To evaluate our model robustly and to get an estimate of radiologist performance, we collect additional labels from board-certified Stanford radiologists on the test set, consisting of 209 musculoskeletal studies. We compared our model and radiologists on the Cohen's kappa statistic, which expresses the agreement of our model and of each radiologist with the gold standard, defined as the majority vote of a disjoint group of radiologists. We find that our model achieves performance comparable to that of radiologists. Model performance is higher than the best radiologist performance in detecting abnormalities on finger studies and equivalent on wrist studies. However, model performance is lower than best radiologist performance in detecting abnormalities on elbow, forearm, hand, humerus, and shoulder studies, indicating that the task is a good challenge for future research. To encourage advances, we have made our dataset freely available at https://stanfordmlgroup.github.io/projects/mura","added-at":"2018-01-19T10:15:17.000+0100","author":[{"propositions":[],"lastnames":["Rajpurkar"],"firstnames":["Pranav"],"suffixes":[]},{"propositions":[],"lastnames":["Irvin"],"firstnames":["Jeremy"],"suffixes":[]},{"propositions":[],"lastnames":["Bagul"],"firstnames":["Aarti"],"suffixes":[]},{"propositions":[],"lastnames":["Ding"],"firstnames":["Daisy"],"suffixes":[]},{"propositions":[],"lastnames":["Duan"],"firstnames":["Tony"],"suffixes":[]},{"propositions":[],"lastnames":["Mehta"],"firstnames":["Hershel"],"suffixes":[]},{"propositions":[],"lastnames":["Yang"],"firstnames":["Brandon"],"suffixes":[]},{"propositions":[],"lastnames":["Zhu"],"firstnames":["Kaylie"],"suffixes":[]},{"propositions":[],"lastnames":["Laird"],"firstnames":["Dillon"],"suffixes":[]},{"propositions":[],"lastnames":["Ball"],"firstnames":["Robyn","L."],"suffixes":[]},{"propositions":[],"lastnames":["Langlotz"],"firstnames":["Curtis"],"suffixes":[]},{"propositions":[],"lastnames":["Shpanskaya"],"firstnames":["Katie"],"suffixes":[]},{"propositions":[],"lastnames":["Lungren"],"firstnames":["Matthew","P."],"suffixes":[]},{"propositions":[],"lastnames":["Ng"],"firstnames":["Andrew"],"suffixes":[]}],"biburl":"https://www.bibsonomy.org/bibtex/2c8ff2f68a0e57716153eefdda755bfd8/alexattia","description":"MURA Dataset: Towards Radiologist-Level Abnormality Detection in Musculoskeletal Radiographs","interhash":"0fbf86d39e14c19e59ea2d81d3feda51","intrahash":"c8ff2f68a0e57716153eefdda755bfd8","keywords":"azmed deep-learning recvis","note":"cite arxiv:1712.06957","timestamp":"2018-01-19T10:15:17.000+0100","title":"MURA Dataset: Towards Radiologist-Level Abnormality Detection in Musculoskeletal Radiographs","url":"http://arxiv.org/abs/1712.06957","year":"2017","bibtex":"@misc{rajpurkar2017dataset,\n abstract = {We introduce MURA, a large dataset of musculoskeletal radiographs containing\r\n40,895 images from 14,982 studies, where each study is manually labeled by\r\nradiologists as either normal or abnormal. On this dataset, we train a\r\n169-layer densely connected convolutional network to detect and localize\r\nabnormalities. To evaluate our model robustly and to get an estimate of\r\nradiologist performance, we collect additional labels from board-certified\r\nStanford radiologists on the test set, consisting of 209 musculoskeletal\r\nstudies. We compared our model and radiologists on the Cohen's kappa statistic,\r\nwhich expresses the agreement of our model and of each radiologist with the\r\ngold standard, defined as the majority vote of a disjoint group of\r\nradiologists. We find that our model achieves performance comparable to that of\r\nradiologists. Model performance is higher than the best radiologist performance\r\nin detecting abnormalities on finger studies and equivalent on wrist studies.\r\nHowever, model performance is lower than best radiologist performance in\r\ndetecting abnormalities on elbow, forearm, hand, humerus, and shoulder studies,\r\nindicating that the task is a good challenge for future research. To encourage\r\nadvances, we have made our dataset freely available at\r\nhttps://stanfordmlgroup.github.io/projects/mura},\n added-at = {2018-01-19T10:15:17.000+0100},\n author = {Rajpurkar, Pranav and Irvin, Jeremy and Bagul, Aarti and Ding, Daisy and Duan, Tony and Mehta, Hershel and Yang, Brandon and Zhu, Kaylie and Laird, Dillon and Ball, Robyn L. and Langlotz, Curtis and Shpanskaya, Katie and Lungren, Matthew P. and Ng, Andrew},\n biburl = {https://www.bibsonomy.org/bibtex/2c8ff2f68a0e57716153eefdda755bfd8/alexattia},\n description = {MURA Dataset: Towards Radiologist-Level Abnormality Detection in\r\n Musculoskeletal Radiographs},\n interhash = {0fbf86d39e14c19e59ea2d81d3feda51},\n intrahash = {c8ff2f68a0e57716153eefdda755bfd8},\n keywords = {azmed deep-learning recvis},\n note = {cite arxiv:1712.06957},\n timestamp = {2018-01-19T10:15:17.000+0100},\n title = {MURA Dataset: Towards Radiologist-Level Abnormality Detection in\r\n Musculoskeletal Radiographs},\n url = {http://arxiv.org/abs/1712.06957},\n year = 2017\n}\n\n","author_short":["Rajpurkar, P.","Irvin, J.","Bagul, A.","Ding, D.","Duan, T.","Mehta, H.","Yang, B.","Zhu, K.","Laird, D.","Ball, R. L.","Langlotz, C.","Shpanskaya, K.","Lungren, M. P.","Ng, A."],"key":"rajpurkar2017dataset","id":"rajpurkar2017dataset","bibbaseid":"rajpurkar-irvin-bagul-ding-duan-mehta-yang-zhu-etal-muradatasettowardsradiologistlevelabnormalitydetectioninmusculoskeletalradiographs-2017","role":"author","urls":{"Paper":"http://arxiv.org/abs/1712.06957"},"keyword":["azmed deep-learning recvis"],"metadata":{"authorlinks":{}}},"bibtype":"misc","biburl":"http://www.bibsonomy.org/bib/author/pranav rajpurkar?items=1000","dataSources":["HKB65fwLfWcKwA8jH","sdazFM6Jfqy88Cwvi"],"keywords":["azmed deep-learning recvis"],"search_terms":["mura","dataset","towards","radiologist","level","abnormality","detection","musculoskeletal","radiographs","rajpurkar","irvin","bagul","ding","duan","mehta","yang","zhu","laird","ball","langlotz","shpanskaya","lungren","ng"],"title":"MURA Dataset: Towards Radiologist-Level Abnormality Detection in Musculoskeletal Radiographs","year":2017}