As the competition winds down, I'd like to take a moment to appreciate all the work that went into this competition. I've worked a lot with medical imaging, and the amount of effort it takes to coordinate a competition like this is tremendous.
\nFirst, collecting the dataset is a huge task in and of itself. Releasing it for public use adds an extra layer of complexity given patient privacy concerns and the need for de-identification. It's awesome to see open data in medical imaging being pushed forward, especially considering that the current incentives to release data publicly and freely are few. Thank you to all the data contributors: AlfredHealth, Koç University Hospital, Stanford University AIMI, Unity Health Toronto, and UNIFESP.
\nAnnotating the data is also very time-consuming. I'm really impressed by the number of radiologists from who volunteered their time to annotate the studies for this challenge. Radiology is busy enough and to spend hours of their valuable time labeling these studies for the greater good is really amazing. A list of annotators are available in the Acknowledgements section of the competition.
\nFinally, coordinating a competition with data across multiple institutions and volunteers from around the world makes things even more challenging. I don't know what goes on in the background, but I imagine that defining a process for these studies to be read as consistently as possible is very difficult. There are also tons of nuances to consider beyond what we see here in the labels. So, a huge thanks to the organizers as well for putting this all together.
\nLooking forward to see all the top solutions in a few days. Best of luck to all the teams!
", "messages": [ { "id": 1058395, "postDate": "2020-10-23T16:47:09.943Z", "content": "As the competition winds down, I'd like to take a moment to appreciate all the work that went into this competition. I've worked a lot with medical imaging, and the amount of effort it takes to coordinate a competition like this is tremendous.
\nFirst, collecting the dataset is a huge task in and of itself. Releasing it for public use adds an extra layer of complexity given patient privacy concerns and the need for de-identification. It's awesome to see open data in medical imaging being pushed forward, especially considering that the current incentives to release data publicly and freely are few. Thank you to all the data contributors: AlfredHealth, Koç University Hospital, Stanford University AIMI, Unity Health Toronto, and UNIFESP.
\nAnnotating the data is also very time-consuming. I'm really impressed by the number of radiologists from who volunteered their time to annotate the studies for this challenge. Radiology is busy enough and to spend hours of their valuable time labeling these studies for the greater good is really amazing. A list of annotators are available in the Acknowledgements section of the competition.
\nFinally, coordinating a competition with data across multiple institutions and volunteers from around the world makes things even more challenging. I don't know what goes on in the background, but I imagine that defining a process for these studies to be read as consistently as possible is very difficult. There are also tons of nuances to consider beyond what we see here in the labels. So, a huge thanks to the organizers as well for putting this all together.
\nLooking forward to see all the top solutions in a few days. Best of luck to all the teams!
", "rawMarkdown": "As the competition winds down, I'd like to take a moment to appreciate all the work that went into this competition. I've worked a lot with medical imaging, and the amount of effort it takes to coordinate a competition like this is tremendous. \n\nFirst, collecting the dataset is a huge task in and of itself. Releasing it for public use adds an extra layer of complexity given patient privacy concerns and the need for de-identification. It's awesome to see open data in medical imaging being pushed forward, especially considering that the current incentives to release data publicly and freely are few. Thank you to all the data contributors: AlfredHealth, Koç University Hospital, Stanford University AIMI, Unity Health Toronto, and UNIFESP. \n\nAnnotating the data is also very time-consuming. I'm really impressed by the number of radiologists from who volunteered their time to annotate the studies for this challenge. Radiology is busy enough and to spend hours of their valuable time labeling these studies for the greater good is really amazing. A list of annotators are available in the Acknowledgements section of the competition.\n\nFinally, coordinating a competition with data across multiple institutions and volunteers from around the world makes things even more challenging. I don't know what goes on in the background, but I imagine that defining a process for these studies to be read as consistently as possible is very difficult. There are also tons of nuances to consider beyond what we see here in the labels. So, a huge thanks to the organizers as well for putting this all together. \n\nLooking forward to see all the top solutions in a few days. Best of luck to all the teams!\n\n", "votes": 46 }, { "id": 1060574, "postDate": "2020-10-26T10:38:36.390Z", "content": "@vaillant ,
\nindeed, the annotation process is very time-consuming. Sadly we started too late after the Lung Fibrosis Competition with this one, nevertheless i managed (with the huge help from @sainatarajan7) to annotate like ~80 RV/LV and sparsely annotate ~ 60 scans with central PE. Both of the datasets i am going to publish soon!
I appreciate all people who spent time to hold this great competition.
\nOf course, I really appreciate you, @vaillant. Without your jpeg dataset, I could not join this competition.
\nI have studied a lot from this competition, thank you very much!
Couldn't agree more. This effort will be worth it. The competition has been a fun and educational experience but in addition, I can see this dataset be used, and reused over and over for research projects, PhD theses, education, etc. It is going to benefit so many people, especially the patients.
\nHomage to Geoff Robinson (aka INcontroL)
", "rawMarkdown": "Couldn't agree more. This effort will be worth it. The competition has been a fun and educational experience but in addition, I can see this dataset be used, and reused over and over for research projects, PhD theses, education, etc. It is going to benefit so many people, especially the patients. \n\nHomage to Geoff Robinson (aka INcontroL)", "votes": 2 }, { "id": 2213302, "postDate": "2023-04-07T13:30:34.783Z", "content": "Hello can somebody tell me about annotations ? I don’t see them anywhere
", "rawMarkdown": "Hello can somebody tell me about annotations ? I don’t see them anywhere " }, { "id": 2213283, "postDate": "2023-04-07T13:23:48.203Z", "content": "Hello @vaillant i am just new in deep learning and on kaggle,
\nwhere did you find annotations of the competition ?
\ni checked on every tab and subtab i didn't saw annotations .
I've learned a lot and as a bonus had a lot of fun. The magnitude of this dataset is amazing, much more so given it's medical images. Thank you @vaillant, your dataset was critical to making it accessible to everyone.
", "rawMarkdown": "I've learned a lot and as a bonus had a lot of fun. The magnitude of this dataset is amazing, much more so given it's medical images. Thank you @vaillant, your dataset was critical to making it accessible to everyone." } ], "comments": [ { "id": 1060574, "author_name": "dr. Konya", "author_url": "", "post_date": "2020-10-26T10:38:36.390000", "content": "@vaillant ,
\nindeed, the annotation process is very time-consuming. Sadly we started too late after the Lung Fibrosis Competition with this one, nevertheless i managed (with the huge help from @sainatarajan7) to annotate like ~80 RV/LV and sparsely annotate ~ 60 scans with central PE. Both of the datasets i am going to publish soon!
I appreciate all people who spent time to hold this great competition.
\nOf course, I really appreciate you, @vaillant. Without your jpeg dataset, I could not join this competition.
\nI have studied a lot from this competition, thank you very much!
Couldn't agree more. This effort will be worth it. The competition has been a fun and educational experience but in addition, I can see this dataset be used, and reused over and over for research projects, PhD theses, education, etc. It is going to benefit so many people, especially the patients.
\nHomage to Geoff Robinson (aka INcontroL)
", "votes": 2, "replies": [] }, { "id": 2213302, "author_name": "steven kombe", "author_url": "", "post_date": "2023-04-07T13:30:34.783000", "content": "Hello can somebody tell me about annotations ? I don’t see them anywhere
", "votes": 0, "replies": [] }, { "id": 2213283, "author_name": "steven kombe", "author_url": "", "post_date": "2023-04-07T13:23:48.203000", "content": "Hello @vaillant i am just new in deep learning and on kaggle,
\nwhere did you find annotations of the competition ?
\ni checked on every tab and subtab i didn't saw annotations .
I've learned a lot and as a bonus had a lot of fun. The magnitude of this dataset is amazing, much more so given it's medical images. Thank you @vaillant, your dataset was critical to making it accessible to everyone.
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1058395": "As the competition winds down, I'd like to take a moment to appreciate all the work that went into this competition. I've worked a lot with medical imaging, and the amount of effort it takes to coordinate a competition like this is tremendous. \n\nFirst, collecting the dataset is a huge task in and of itself. Releasing it for public use adds an extra layer of complexity given patient privacy concerns and the need for de-identification. It's awesome to see open data in medical imaging being pushed forward, especially considering that the current incentives to release data publicly and freely are few. Thank you to all the data contributors: AlfredHealth, Koç University Hospital, Stanford University AIMI, Unity Health Toronto, and UNIFESP. \n\nAnnotating the data is also very time-consuming. I'm really impressed by the number of radiologists from who volunteered their time to annotate the studies for this challenge. Radiology is busy enough and to spend hours of their valuable time labeling these studies for the greater good is really amazing. A list of annotators are available in the Acknowledgements section of the competition.\n\nFinally, coordinating a competition with data across multiple institutions and volunteers from around the world makes things even more challenging. I don't know what goes on in the background, but I imagine that defining a process for these studies to be read as consistently as possible is very difficult. There are also tons of nuances to consider beyond what we see here in the labels. So, a huge thanks to the organizers as well for putting this all together. \n\nLooking forward to see all the top solutions in a few days. Best of luck to all the teams!\n\n", "1060574": "@vaillant ,\nindeed, the annotation process is very time-consuming. Sadly we started too late after the Lung Fibrosis Competition with this one, nevertheless i managed (with the huge help from @sainatarajan7) to annotate like ~80 RV/LV and sparsely annotate ~ 60 scans with central PE. Both of the datasets i am going to publish soon!", "1058420": "I appreciate all people who spent time to hold this great competition. \nOf course, I really appreciate you, @vaillant. Without your jpeg dataset, I could not join this competition. \nI have studied a lot from this competition, thank you very much!", "1058443": "Couldn't agree more. This effort will be worth it. The competition has been a fun and educational experience but in addition, I can see this dataset be used, and reused over and over for research projects, PhD theses, education, etc. It is going to benefit so many people, especially the patients. \n\nHomage to Geoff Robinson (aka INcontroL)", "2213302": "Hello can somebody tell me about annotations ? I don’t see them anywhere ", "2213283": "Hello @vaillant i am just new in deep learning and on kaggle, \nwhere did you find annotations of the competition ?\ni checked on every tab and subtab i didn't saw annotations . \n", "1058659": "I've learned a lot and as a bonus had a lot of fun. The magnitude of this dataset is amazing, much more so given it's medical images. Thank you @vaillant, your dataset was critical to making it accessible to everyone." } }