{
  "id": 182809,
  "title": "FUMPE dataset made available for everyone",
  "url": "/competitions/rsna-str-pulmonary-embolism-detection/discussion/182809",
  "author_name": "Larxel",
  "post_date": "2020-09-14T12:09:30.491000",
  "votes": 24,
  "comment_count": 6,
  "views": 0,
  "content": "<p>Hello frens</p>\n<p>A couple of months ago I uploaded a <strong><a href=\"https://www.kaggle.com/andrewmvd/pulmonary-embolism-in-ct-images\" target=\"_blank\">Pulmonary Embolism segmentation dataset</a></strong> called FUMPE (standing for Ferdowsi University of Mashhad's Pulmonary Embolism dataset), and it is highly likely that it will be useful here.<br>\nPersonaly, I'm happy that there is a competition about PE as it is such a relevant issue worldwide and providing these datasets is a way that I found to help the overall results.</p>\n<p>FUMPE does not share the same labels, yet it may prove invaluable for this competition.<br>\nIn FUMPE, the data is for segmentation purposes, which could be used to further refine the classification task in this competition, as well as transfer learning from the same imaging domain (CTA).</p>\n<p>The authors of the dataset are:</p>\n<blockquote>\n  <p>Masoudi, M. et al. A new dataset of computed-tomography angiography images for computer-aided detection of pulmonary embolism. Sci. Data 5:180180 doi: 10.1038/sdata.2018.178 (2018).</p>\n</blockquote>\n<p>You can view their article published on nature <a href=\"https://www.nature.com/articles/sdata2018180\" target=\"_blank\">here</a>.</p>\n<p>As always, if you ever publish anything, please credit the authors.</p>\n<p>Hope it helps<br>\nHave a nice competition!</p>",
  "messages": [
    {
      "id": 1009965,
      "postDate": "2020-09-14T12:09:30.493Z",
      "content": "<p>Hello frens</p>\n<p>A couple of months ago I uploaded a <strong><a href=\"https://www.kaggle.com/andrewmvd/pulmonary-embolism-in-ct-images\" target=\"_blank\">Pulmonary Embolism segmentation dataset</a></strong> called FUMPE (standing for Ferdowsi University of Mashhad's Pulmonary Embolism dataset), and it is highly likely that it will be useful here.<br>\nPersonaly, I'm happy that there is a competition about PE as it is such a relevant issue worldwide and providing these datasets is a way that I found to help the overall results.</p>\n<p>FUMPE does not share the same labels, yet it may prove invaluable for this competition.<br>\nIn FUMPE, the data is for segmentation purposes, which could be used to further refine the classification task in this competition, as well as transfer learning from the same imaging domain (CTA).</p>\n<p>The authors of the dataset are:</p>\n<blockquote>\n  <p>Masoudi, M. et al. A new dataset of computed-tomography angiography images for computer-aided detection of pulmonary embolism. Sci. Data 5:180180 doi: 10.1038/sdata.2018.178 (2018).</p>\n</blockquote>\n<p>You can view their article published on nature <a href=\"https://www.nature.com/articles/sdata2018180\" target=\"_blank\">here</a>.</p>\n<p>As always, if you ever publish anything, please credit the authors.</p>\n<p>Hope it helps<br>\nHave a nice competition!</p>",
      "rawMarkdown": "Hello frens\n\nA couple of months ago I uploaded a **[Pulmonary Embolism segmentation dataset](https://www.kaggle.com/andrewmvd/pulmonary-embolism-in-ct-images)** called FUMPE (standing for Ferdowsi University of Mashhad's Pulmonary Embolism dataset), and it is highly likely that it will be useful here.\nPersonaly, I'm happy that there is a competition about PE as it is such a relevant issue worldwide and providing these datasets is a way that I found to help the overall results.\n\nFUMPE does not share the same labels, yet it may prove invaluable for this competition.\nIn FUMPE, the data is for segmentation purposes, which could be used to further refine the classification task in this competition, as well as transfer learning from the same imaging domain (CTA).\n\nThe authors of the dataset are:\n> Masoudi, M. et al. A new dataset of computed-tomography angiography images for computer-aided detection of pulmonary embolism. Sci. Data 5:180180 doi: 10.1038/sdata.2018.178 (2018).\n\nYou can view their article published on nature [here](https://www.nature.com/articles/sdata2018180).\n\nAs always, if you ever publish anything, please credit the authors.\n\nHope it helps\nHave a nice competition!",
      "votes": 24
    },
    {
      "id": 1015983,
      "postDate": "2020-09-18T15:22:40.903Z",
      "content": "<p>Thanks a lot for sharing. I will have a look on that dataset for sure.!!!</p>",
      "rawMarkdown": "Thanks a lot for sharing. I will have a look on that dataset for sure.!!!",
      "votes": 1
    },
    {
      "id": 1058644,
      "postDate": "2020-10-24T03:08:38.607Z",
      "content": "<p>It would be useful to train a combined self-supervised model to further do transfer learning. Thanks for sharing!</p>",
      "rawMarkdown": "It would be useful to train a combined self-supervised model to further do transfer learning. Thanks for sharing!"
    },
    {
      "id": 1020153,
      "postDate": "2020-09-21T01:34:25.390Z",
      "content": "<p><a href=\"https://www.kaggle.com/philculliton\" target=\"_blank\">@philculliton</a> </p>\n<p>What is the competition deadline for posting links to any external data used? <br>\nCan we label features in a subset of the training data and post them here as an additional dataset?</p>",
      "rawMarkdown": "@philculliton \n\nWhat is the competition deadline for posting links to any external data used? \nCan we label features in a subset of the training data and post them here as an additional dataset?"
    },
    {
      "id": 1017857,
      "postDate": "2020-09-19T09:55:00.387Z",
      "content": "<p>Dear <a href=\"https://www.kaggle.com/andrewmvd\" target=\"_blank\">@andrewmvd</a>, I am currently working on a this dataset. But finding it a  bit hard to use 3D convolutions. Can you share or direct me to some of the resources regarding 3D convolutions and such..</p>\n<p>Thanks.. </p>",
      "rawMarkdown": "Dear @andrewmvd, I am currently working on a this dataset. But finding it a  bit hard to use 3D convolutions. Can you share or direct me to some of the resources regarding 3D convolutions and such..\n\nThanks.. \n"
    },
    {
      "id": 1020067,
      "postDate": "2020-09-20T21:49:55.233Z",
      "content": "<p>Thanks for sharing this!</p>",
      "rawMarkdown": "Thanks for sharing this!",
      "votes": 1
    },
    {
      "id": 1013109,
      "postDate": "2020-09-16T14:03:50.393Z",
      "content": "<p>Thanks a lot for sharing!</p>",
      "rawMarkdown": "Thanks a lot for sharing!",
      "votes": 1
    }
  ],
  "comments": [
    {
      "id": 1015983,
      "author_name": "Redwan Sony",
      "author_url": "",
      "post_date": "2020-09-18T15:22:40.903000",
      "content": "<p>Thanks a lot for sharing. I will have a look on that dataset for sure.!!!</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 1058644,
      "author_name": "Kerem Turgutlu",
      "author_url": "",
      "post_date": "2020-10-24T03:08:38.607000",
      "content": "<p>It would be useful to train a combined self-supervised model to further do transfer learning. Thanks for sharing!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1020153,
      "author_name": "aksg87",
      "author_url": "",
      "post_date": "2020-09-21T01:34:25.390000",
      "content": "<p><a href=\"https://www.kaggle.com/philculliton\" target=\"_blank\">@philculliton</a> </p>\n<p>What is the competition deadline for posting links to any external data used? <br>\nCan we label features in a subset of the training data and post them here as an additional dataset?</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1017857,
      "author_name": "Redwan Sony",
      "author_url": "",
      "post_date": "2020-09-19T09:55:00.387000",
      "content": "<p>Dear <a href=\"https://www.kaggle.com/andrewmvd\" target=\"_blank\">@andrewmvd</a>, I am currently working on a this dataset. But finding it a  bit hard to use 3D convolutions. Can you share or direct me to some of the resources regarding 3D convolutions and such..</p>\n<p>Thanks.. </p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 1020067,
      "author_name": "aksg87",
      "author_url": "",
      "post_date": "2020-09-20T21:49:55.233000",
      "content": "<p>Thanks for sharing this!</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 1013109,
      "author_name": "Amritvir Singh",
      "author_url": "",
      "post_date": "2020-09-16T14:03:50.393000",
      "content": "<p>Thanks a lot for sharing!</p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1009965": "Hello frens\n\nA couple of months ago I uploaded a **[Pulmonary Embolism segmentation dataset](https://www.kaggle.com/andrewmvd/pulmonary-embolism-in-ct-images)** called FUMPE (standing for Ferdowsi University of Mashhad's Pulmonary Embolism dataset), and it is highly likely that it will be useful here.\nPersonaly, I'm happy that there is a competition about PE as it is such a relevant issue worldwide and providing these datasets is a way that I found to help the overall results.\n\nFUMPE does not share the same labels, yet it may prove invaluable for this competition.\nIn FUMPE, the data is for segmentation purposes, which could be used to further refine the classification task in this competition, as well as transfer learning from the same imaging domain (CTA).\n\nThe authors of the dataset are:\n> Masoudi, M. et al. A new dataset of computed-tomography angiography images for computer-aided detection of pulmonary embolism. Sci. Data 5:180180 doi: 10.1038/sdata.2018.178 (2018).\n\nYou can view their article published on nature [here](https://www.nature.com/articles/sdata2018180).\n\nAs always, if you ever publish anything, please credit the authors.\n\nHope it helps\nHave a nice competition!",
    "1015983": "Thanks a lot for sharing. I will have a look on that dataset for sure.!!!",
    "1058644": "It would be useful to train a combined self-supervised model to further do transfer learning. Thanks for sharing!",
    "1020153": "@philculliton \n\nWhat is the competition deadline for posting links to any external data used? \nCan we label features in a subset of the training data and post them here as an additional dataset?",
    "1017857": "Dear @andrewmvd, I am currently working on a this dataset. But finding it a  bit hard to use 3D convolutions. Can you share or direct me to some of the resources regarding 3D convolutions and such..\n\nThanks.. \n",
    "1020067": "Thanks for sharing this!",
    "1013109": "Thanks a lot for sharing!"
  }
}