Being myself a radiologist, I can definitely understand how it can be tedious to annotate several thousand xray images with bounding boxes.
\n\nTo my knowledge, the current challenge dataset is probably the largest publicly available dataset with strong annotations and bounding boxes on 2D images in radiology. I am sure that this dataset can generate many new research ideas beyond the current competition.
\n\nEthically, creating any public dataset in the healthcare field is very challenging. But, it is most probably one of the best way to advance the field forward. I really encourage international healthcare institutions and research labs to continue to pave the way in releasing strongly deidentified public datasets.
\n\nConsequently, I had today a truly respectul thought to all the annotators/organizers that allowed us to train models on this subset of the original NIH dataset published by Ron Summers. Kudos!
\n\nhttps://www.kaggle.com/c/rsna-pneumonia-detection-challenge#Acknowledgements
", "messages": [ { "id": "408298", "postDate": "10/22/2018 16:15:57", "content": "Being myself a radiologist, I can definitely understand how it can be tedious to annotate several thousand xray images with bounding boxes.
\n\nTo my knowledge, the current challenge dataset is probably the largest publicly available dataset with strong annotations and bounding boxes on 2D images in radiology. I am sure that this dataset can generate many new research ideas beyond the current competition.
\n\nEthically, creating any public dataset in the healthcare field is very challenging. But, it is most probably one of the best way to advance the field forward. I really encourage international healthcare institutions and research labs to continue to pave the way in releasing strongly deidentified public datasets.
\n\nConsequently, I had today a truly respectul thought to all the annotators/organizers that allowed us to train models on this subset of the original NIH dataset published by Ron Summers. Kudos!
\n\nhttps://www.kaggle.com/c/rsna-pneumonia-detection-challenge#Acknowledgements
", "rawMarkdown": "Being myself a radiologist, I can definitely understand how it can be tedious to annotate several thousand xray images with bounding boxes.\n\nTo my knowledge, the current challenge dataset is probably the largest publicly available dataset with strong annotations and bounding boxes on 2D images in radiology. I am sure that this dataset can generate many new research ideas beyond the current competition.\n\nEthically, creating any public dataset in the healthcare field is very challenging. But, it is most probably one of the best way to advance the field forward. I really encourage international healthcare institutions and research labs to continue to pave the way in releasing strongly deidentified public datasets.\n\nConsequently, I had today a truly respectul thought to all the annotators/organizers that allowed us to train models on this subset of the original NIH dataset published by Ron Summers. Kudos!\n\nhttps://www.kaggle.com/c/rsna-pneumonia-detection-challenge#Acknowledgements", "votes": null }, { "id": "408320", "postDate": "10/22/2018 17:09:39", "content": "Agreed, major respect to the annotators of this amazingly high quality data set! Well done. Can't wait to see what approaches people have taken and the performance on the final test set. It's been exciting to be part of such a well organised and important competition, so thank you.
\n\nJust in case anyone misses it on the Overview tab above, here are the many acknowledgements for this competition:
\n\n\n", "rawMarkdown": "Agreed, major respect to the annotators of this amazingly high quality data set! Well done. Can't wait to see what approaches people have taken and the performance on the final test set. It's been exciting to be part of such a well organised and important competition, so thank you.\n\nJust in case anyone misses it on the Overview tab above, here are the many acknowledgements for this competition:\n\n\n>Kaggle and RSNA would like to recognize the radiologists from the Society of Thoracic Radiology and RSNA who contributed the considerable effort required to annotate the datasets in preparation for the challenge:\n\n>Society of Thoracic Radiology \n\n>Judith Amorosa, MD- Rutgers Robert Wood Johnson Medical School \n\n>Veronica Arteaga, MD - University of Arizona * \n\n>Maya Galperin-Aizenberg, MD - University of Pennsylvania \n\n>Ritu Gill, MD - Beth Israel Deaconess Medical Center * \n\n>Myrna Godoy, MD, PhD - MD Anderson Cancer Center * \n\n>Stephen Hobbs, MD - University of Kentucky * \n\n>Jean Jeudy, MD - University of Maryland * \n\n>Archana Laroia, MD - University of Iowa * \n\n>Palmi Shah, MD - Rush University * \n\n>Dharshan Vummidi, MD - University of Michigan * \n\n>Carol Wu, MD - MD Anderson Cancer Center *\n\n>Kavitha Yaddanapudi, MD - Stony Brook School of Medicine * \n\n>Radiological Society of North America\n\n>Tessa Cook, MD, PhD - University of Pennsylvania *\n\n>Safwan Halabi, MD - Stanford University *\n\n>Marc Kohli, MD - University of California - San Francisco *\n\n>Luciano M Prevedello, MD, MPH - The Ohio State University *\n\n>Arjun Sharma, MD - Amita Health *\n\n>George Shih, MD - Weill Cornell Medicine *\n\n>* 1,000 cases or more\n\n>Special thanks to Anouk Stein, MD of MD.ai, who contributed the annotation tools used in creating the challenge datasets, helped to organize the team of annotators and provided extensive consulting on the annotation process.\n\n>Special thanks to Jayashree Kalpathy-Cramer, PhD, Massachusetts General Hospital; Peter Chang, MD and John Mongan, MD, PhD, University of California - San Francisco; and George Shih, MD, Weill Cornell Medicine for their work in validating the evaluation metric used in the challenge.\n\n>Finally, thanks to the members of the RSNA Machine Learning Steering and Machine Learning Data Standards subcommittees, who were responsible for designing, organizing and conducting the challenge in collaboration with Kaggle:\n\n>Katherine P. Andriole, PhD - MGH & BWH Center for Clinical Data Science\n\n>Falgun H. Chokshi, MD - Emory University\n\n>Brad Erickson, MD - Mayo Clinic\n\n>Adam Flanders, MD - Thomas Jefferson University\n\n>Safwan Halabi, MD - Stanford University\n\n>Jayashree Kalpathy-Cramer, PhD - Massachusetts General Hospital\n\n>Marc Kohli, University of California, MD - San Francisco\n\n>Luciano Prevedello, MD, MPH - The Ohio State University\n\n>George Shih, MD - Weill Cornell Medicine\n\n>Carol Wu, MD - MD Anderson Cancer Center", "votes": null }, { "id": "408547", "postDate": "10/23/2018 04:10:24", "content": "Kaggle and RSNA would like to recognize the radiologists from the Society of Thoracic Radiology and RSNA who contributed the considerable effort required to annotate the datasets in preparation for the challenge:
\n \nSociety of Thoracic Radiology
\n \nJudith Amorosa, MD- Rutgers Robert Wood Johnson Medical School
\n \nVeronica Arteaga, MD - University of Arizona *
\n \nMaya Galperin-Aizenberg, MD - University of Pennsylvania
\n \nRitu Gill, MD - Beth Israel Deaconess Medical Center *
\n \nMyrna Godoy, MD, PhD - MD Anderson Cancer Center *
\n \nStephen Hobbs, MD - University of Kentucky *
\n \nJean Jeudy, MD - University of Maryland *
\n \nArchana Laroia, MD - University of Iowa *
\n \nPalmi Shah, MD - Rush University *
\n \nDharshan Vummidi, MD - University of Michigan *
\n \nCarol Wu, MD - MD Anderson Cancer Center *
\n \nKavitha Yaddanapudi, MD - Stony Brook School of Medicine *
\n \nRadiological Society of North America
\n \nTessa Cook, MD, PhD - University of Pennsylvania *
\n \nSafwan Halabi, MD - Stanford University *
\n \nMarc Kohli, MD - University of California - San Francisco *
\n \nLuciano M Prevedello, MD, MPH - The Ohio State University *
\n \nArjun Sharma, MD - Amita Health *
\n \nGeorge Shih, MD - Weill Cornell Medicine *
\n \n\n
\n \n- 1,000 cases or more
\nSpecial thanks to Anouk Stein, MD of MD.ai, who contributed the annotation tools used in creating the challenge datasets, helped to organize the team of annotators and provided extensive consulting on the annotation process.
\n \nSpecial thanks to Jayashree Kalpathy-Cramer, PhD, Massachusetts General Hospital; Peter Chang, MD and John Mongan, MD, PhD, University of California - San Francisco; and George Shih, MD, Weill Cornell Medicine for their work in validating the evaluation metric used in the challenge.
\n \nFinally, thanks to the members of the RSNA Machine Learning Steering and Machine Learning Data Standards subcommittees, who were responsible for designing, organizing and conducting the challenge in collaboration with Kaggle:
\n \nKatherine P. Andriole, PhD - MGH & BWH Center for Clinical Data Science
\n \nFalgun H. Chokshi, MD - Emory University
\n \nBrad Erickson, MD - Mayo Clinic
\n \nAdam Flanders, MD - Thomas Jefferson University
\n \nSafwan Halabi, MD - Stanford University
\n \nJayashree Kalpathy-Cramer, PhD - Massachusetts General Hospital
\n \nMarc Kohli, University of California, MD - San Francisco
\n \nLuciano Prevedello, MD, MPH - The Ohio State University
\n \nGeorge Shih, MD - Weill Cornell Medicine
\n \nCarol Wu, MD - MD Anderson Cancer Center
\n
I also would like to also voice my support to the radiologists that annotated the dataset, and invariably reviewed the edge cases for consensus and accuracy for inclusion in this challenge. I am sure the annotators wish never to see another 'fluffy opacity' anytime soon. I am also sure the friends and family of those participating in the challenge will be happy the challenge is ending soon. Perhaps the cloud service providers will shed a tear for lost revenues as we resume more 'normal' activities.
\n\nThe contribution of organized radiology to the advancement of science is evident in this project, and similar ones spearheaded by the radiology community. Radiologists should feel proud of participating in these forward-thinking efforts, and encourage colleagues' participation in similar future efforts.
", "rawMarkdown": "I also would like to also voice my support to the radiologists that annotated the dataset, and invariably reviewed the edge cases for consensus and accuracy for inclusion in this challenge. I am sure the annotators wish never to see another 'fluffy opacity' anytime soon. I am also sure the friends and family of those participating in the challenge will be happy the challenge is ending soon. Perhaps the cloud service providers will shed a tear for lost revenues as we resume more 'normal' activities. \n\nThe contribution of organized radiology to the advancement of science is evident in this project, and similar ones spearheaded by the radiology community. Radiologists should feel proud of participating in these forward-thinking efforts, and encourage colleagues' participation in similar future efforts.", "votes": null } ], "comments": [ { "id": 408320, "author_name": "taindow", "author_url": "", "post_date": "10/22/2018 17:09:39", "content": "Agreed, major respect to the annotators of this amazingly high quality data set! Well done. Can't wait to see what approaches people have taken and the performance on the final test set. It's been exciting to be part of such a well organised and important competition, so thank you.
\n\nJust in case anyone misses it on the Overview tab above, here are the many acknowledgements for this competition:
\n\n\n", "votes": null, "replies": [] }, { "id": 408547, "author_name": "drsxr1", "author_url": "", "post_date": "10/23/2018 04:10:24", "content": "Kaggle and RSNA would like to recognize the radiologists from the Society of Thoracic Radiology and RSNA who contributed the considerable effort required to annotate the datasets in preparation for the challenge:
\n \nSociety of Thoracic Radiology
\n \nJudith Amorosa, MD- Rutgers Robert Wood Johnson Medical School
\n \nVeronica Arteaga, MD - University of Arizona *
\n \nMaya Galperin-Aizenberg, MD - University of Pennsylvania
\n \nRitu Gill, MD - Beth Israel Deaconess Medical Center *
\n \nMyrna Godoy, MD, PhD - MD Anderson Cancer Center *
\n \nStephen Hobbs, MD - University of Kentucky *
\n \nJean Jeudy, MD - University of Maryland *
\n \nArchana Laroia, MD - University of Iowa *
\n \nPalmi Shah, MD - Rush University *
\n \nDharshan Vummidi, MD - University of Michigan *
\n \nCarol Wu, MD - MD Anderson Cancer Center *
\n \nKavitha Yaddanapudi, MD - Stony Brook School of Medicine *
\n \nRadiological Society of North America
\n \nTessa Cook, MD, PhD - University of Pennsylvania *
\n \nSafwan Halabi, MD - Stanford University *
\n \nMarc Kohli, MD - University of California - San Francisco *
\n \nLuciano M Prevedello, MD, MPH - The Ohio State University *
\n \nArjun Sharma, MD - Amita Health *
\n \nGeorge Shih, MD - Weill Cornell Medicine *
\n \n\n
\n \n- 1,000 cases or more
\nSpecial thanks to Anouk Stein, MD of MD.ai, who contributed the annotation tools used in creating the challenge datasets, helped to organize the team of annotators and provided extensive consulting on the annotation process.
\n \nSpecial thanks to Jayashree Kalpathy-Cramer, PhD, Massachusetts General Hospital; Peter Chang, MD and John Mongan, MD, PhD, University of California - San Francisco; and George Shih, MD, Weill Cornell Medicine for their work in validating the evaluation metric used in the challenge.
\n \nFinally, thanks to the members of the RSNA Machine Learning Steering and Machine Learning Data Standards subcommittees, who were responsible for designing, organizing and conducting the challenge in collaboration with Kaggle:
\n \nKatherine P. Andriole, PhD - MGH & BWH Center for Clinical Data Science
\n \nFalgun H. Chokshi, MD - Emory University
\n \nBrad Erickson, MD - Mayo Clinic
\n \nAdam Flanders, MD - Thomas Jefferson University
\n \nSafwan Halabi, MD - Stanford University
\n \nJayashree Kalpathy-Cramer, PhD - Massachusetts General Hospital
\n \nMarc Kohli, University of California, MD - San Francisco
\n \nLuciano Prevedello, MD, MPH - The Ohio State University
\n \nGeorge Shih, MD - Weill Cornell Medicine
\n \nCarol Wu, MD - MD Anderson Cancer Center
\n
I also would like to also voice my support to the radiologists that annotated the dataset, and invariably reviewed the edge cases for consensus and accuracy for inclusion in this challenge. I am sure the annotators wish never to see another 'fluffy opacity' anytime soon. I am also sure the friends and family of those participating in the challenge will be happy the challenge is ending soon. Perhaps the cloud service providers will shed a tear for lost revenues as we resume more 'normal' activities.
\n\nThe contribution of organized radiology to the advancement of science is evident in this project, and similar ones spearheaded by the radiology community. Radiologists should feel proud of participating in these forward-thinking efforts, and encourage colleagues' participation in similar future efforts.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "408298": "Being myself a radiologist, I can definitely understand how it can be tedious to annotate several thousand xray images with bounding boxes.\n\nTo my knowledge, the current challenge dataset is probably the largest publicly available dataset with strong annotations and bounding boxes on 2D images in radiology. I am sure that this dataset can generate many new research ideas beyond the current competition.\n\nEthically, creating any public dataset in the healthcare field is very challenging. But, it is most probably one of the best way to advance the field forward. I really encourage international healthcare institutions and research labs to continue to pave the way in releasing strongly deidentified public datasets.\n\nConsequently, I had today a truly respectul thought to all the annotators/organizers that allowed us to train models on this subset of the original NIH dataset published by Ron Summers. Kudos!\n\nhttps://www.kaggle.com/c/rsna-pneumonia-detection-challenge#Acknowledgements", "408320": "Agreed, major respect to the annotators of this amazingly high quality data set! Well done. Can't wait to see what approaches people have taken and the performance on the final test set. It's been exciting to be part of such a well organised and important competition, so thank you.\n\nJust in case anyone misses it on the Overview tab above, here are the many acknowledgements for this competition:\n\n\n>Kaggle and RSNA would like to recognize the radiologists from the Society of Thoracic Radiology and RSNA who contributed the considerable effort required to annotate the datasets in preparation for the challenge:\n\n>Society of Thoracic Radiology \n\n>Judith Amorosa, MD- Rutgers Robert Wood Johnson Medical School \n\n>Veronica Arteaga, MD - University of Arizona * \n\n>Maya Galperin-Aizenberg, MD - University of Pennsylvania \n\n>Ritu Gill, MD - Beth Israel Deaconess Medical Center * \n\n>Myrna Godoy, MD, PhD - MD Anderson Cancer Center * \n\n>Stephen Hobbs, MD - University of Kentucky * \n\n>Jean Jeudy, MD - University of Maryland * \n\n>Archana Laroia, MD - University of Iowa * \n\n>Palmi Shah, MD - Rush University * \n\n>Dharshan Vummidi, MD - University of Michigan * \n\n>Carol Wu, MD - MD Anderson Cancer Center *\n\n>Kavitha Yaddanapudi, MD - Stony Brook School of Medicine * \n\n>Radiological Society of North America\n\n>Tessa Cook, MD, PhD - University of Pennsylvania *\n\n>Safwan Halabi, MD - Stanford University *\n\n>Marc Kohli, MD - University of California - San Francisco *\n\n>Luciano M Prevedello, MD, MPH - The Ohio State University *\n\n>Arjun Sharma, MD - Amita Health *\n\n>George Shih, MD - Weill Cornell Medicine *\n\n>* 1,000 cases or more\n\n>Special thanks to Anouk Stein, MD of MD.ai, who contributed the annotation tools used in creating the challenge datasets, helped to organize the team of annotators and provided extensive consulting on the annotation process.\n\n>Special thanks to Jayashree Kalpathy-Cramer, PhD, Massachusetts General Hospital; Peter Chang, MD and John Mongan, MD, PhD, University of California - San Francisco; and George Shih, MD, Weill Cornell Medicine for their work in validating the evaluation metric used in the challenge.\n\n>Finally, thanks to the members of the RSNA Machine Learning Steering and Machine Learning Data Standards subcommittees, who were responsible for designing, organizing and conducting the challenge in collaboration with Kaggle:\n\n>Katherine P. Andriole, PhD - MGH & BWH Center for Clinical Data Science\n\n>Falgun H. Chokshi, MD - Emory University\n\n>Brad Erickson, MD - Mayo Clinic\n\n>Adam Flanders, MD - Thomas Jefferson University\n\n>Safwan Halabi, MD - Stanford University\n\n>Jayashree Kalpathy-Cramer, PhD - Massachusetts General Hospital\n\n>Marc Kohli, University of California, MD - San Francisco\n\n>Luciano Prevedello, MD, MPH - The Ohio State University\n\n>George Shih, MD - Weill Cornell Medicine\n\n>Carol Wu, MD - MD Anderson Cancer Center", "408547": "I also would like to also voice my support to the radiologists that annotated the dataset, and invariably reviewed the edge cases for consensus and accuracy for inclusion in this challenge. I am sure the annotators wish never to see another 'fluffy opacity' anytime soon. I am also sure the friends and family of those participating in the challenge will be happy the challenge is ending soon. Perhaps the cloud service providers will shed a tear for lost revenues as we resume more 'normal' activities. \n\nThe contribution of organized radiology to the advancement of science is evident in this project, and similar ones spearheaded by the radiology community. Radiologists should feel proud of participating in these forward-thinking efforts, and encourage colleagues' participation in similar future efforts." }, "source": "meta" }