{
  "id": 263956,
  "title": "External Datasets",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/263956",
  "author_name": "",
  "post_date": "2021-08-10T16:27:02.733715Z",
  "votes": 4,
  "comment_count": 6,
  "views": 0,
  "content": "<p>What external datasets do you guys think may be relevant to this competition in terms of additional training data?</p>\n<p>Thanks,</p>",
  "messages": [
    {
      "id": "1464615",
      "postDate": "08/10/2021 16:27:02",
      "content": "<p>What external datasets do you guys think may be relevant to this competition in terms of additional training data?</p>\n<p>Thanks,</p>",
      "rawMarkdown": "What external datasets do you guys think may be relevant to this competition in terms of additional training data?\n\nThanks,",
      "votes": null
    },
    {
      "id": "1465163",
      "postDate": "08/10/2021 22:30:07",
      "content": "<p>I found this one: <a href=\"https://www.kaggle.com/sartajbhuvaji/brain-tumor-classification-mri\" target=\"_blank\">Brain Tumor Classification (MRI)</a><br>\nIt has 4 types of images, one type is no tumor, and another type is glioblastoma.<br>\nBut the description says <code>you will predict the genetic subtype of glioblastoma using MRI (magnetic resonance imaging) scans</code>, I'm not sure if this glioma_tumor is as same as the glioblastoma tumor for this competition. <br>\nThe external dataset combined T1, T2, FLAIR images as well, but you can try it.</p>",
      "rawMarkdown": "I found this one: [Brain Tumor Classification (MRI)](https://www.kaggle.com/sartajbhuvaji/brain-tumor-classification-mri)\nIt has 4 types of images, one type is no tumor, and another type is glioblastoma.\nBut the description says `you will predict the genetic subtype of glioblastoma using MRI (magnetic resonance imaging) scans`, I'm not sure if this glioma_tumor is as same as the glioblastoma tumor for this competition. \nThe external dataset combined T1, T2, FLAIR images as well, but you can try it.",
      "votes": null
    },
    {
      "id": "1465310",
      "postDate": "08/11/2021 02:08:33",
      "content": "<p>Thanks for sharing!!! I think these data are good for training model </p>",
      "rawMarkdown": "Thanks for sharing!!! I think these data are good for training model",
      "votes": null
    },
    {
      "id": "1470601",
      "postDate": "08/13/2021 16:05:30",
      "content": "<p><a href=\"https://www.kaggle.com/adheshgarg/brats12\" target=\"_blank\">https://www.kaggle.com/adheshgarg/brats12</a></p>\n<p>thanks to <a href=\"https://www.kaggle.com/c/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/262120\" target=\"_blank\">https://www.kaggle.com/c/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/262120</a></p>",
      "rawMarkdown": "https://www.kaggle.com/adheshgarg/brats12\n\nthanks to https://www.kaggle.com/c/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/262120",
      "votes": null
    },
    {
      "id": "1470766",
      "postDate": "08/13/2021 17:45:42",
      "content": "<p>I have viewed these images before. I thought they are too blurred to use. According to my experience, low-quality images sometimes damage the model's performance. Did this data improve your model? </p>",
      "rawMarkdown": "I have viewed these images before. I thought they are too blurred to use. According to my experience, low-quality images sometimes damage the model's performance. Did this data improve your model?",
      "votes": null
    },
    {
      "id": "1470780",
      "postDate": "08/13/2021 17:55:15",
      "content": "<p>yet to use😅</p>",
      "rawMarkdown": "yet to use😅",
      "votes": null
    },
    {
      "id": "1471621",
      "postDate": "08/14/2021 10:43:12",
      "content": "<p>There is a separate dataset for the segmentation part of the challenge (Task-1) here: <a href=\"https://www.synapse.org/#!Synapse:syn25829067/wiki/610863\" target=\"_blank\">https://www.synapse.org/#!Synapse:syn25829067/wiki/610863</a></p>\n<p>As far as I understand it, it‘s the same brains in nifti format, but with ground truth segmentation. I did not register for downloading it yet, but think it would be helpful. It would be ok to use that data for the classification part (Task-2) as well, right?</p>",
      "rawMarkdown": "There is a separate dataset for the segmentation part of the challenge (Task-1) here: https://www.synapse.org/#!Synapse:syn25829067/wiki/610863\n\nAs far as I understand it, it‘s the same brains in nifti format, but with ground truth segmentation. I did not register for downloading it yet, but think it would be helpful. It would be ok to use that data for the classification part (Task-2) as well, right?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1465163,
      "author_name": "bcghost",
      "author_url": "",
      "post_date": "08/10/2021 22:30:07",
      "content": "<p>I found this one: <a href=\"https://www.kaggle.com/sartajbhuvaji/brain-tumor-classification-mri\" target=\"_blank\">Brain Tumor Classification (MRI)</a><br>\nIt has 4 types of images, one type is no tumor, and another type is glioblastoma.<br>\nBut the description says <code>you will predict the genetic subtype of glioblastoma using MRI (magnetic resonance imaging) scans</code>, I'm not sure if this glioma_tumor is as same as the glioblastoma tumor for this competition. <br>\nThe external dataset combined T1, T2, FLAIR images as well, but you can try it.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1465310,
          "author_name": "kevinleekrus",
          "author_url": "",
          "post_date": "08/11/2021 02:08:33",
          "content": "<p>Thanks for sharing!!! I think these data are good for training model </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1470601,
      "author_name": "mrinath",
      "author_url": "",
      "post_date": "08/13/2021 16:05:30",
      "content": "<p><a href=\"https://www.kaggle.com/adheshgarg/brats12\" target=\"_blank\">https://www.kaggle.com/adheshgarg/brats12</a></p>\n<p>thanks to <a href=\"https://www.kaggle.com/c/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/262120\" target=\"_blank\">https://www.kaggle.com/c/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/262120</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 1470766,
          "author_name": "bcghost",
          "author_url": "",
          "post_date": "08/13/2021 17:45:42",
          "content": "<p>I have viewed these images before. I thought they are too blurred to use. According to my experience, low-quality images sometimes damage the model's performance. Did this data improve your model? </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1470780,
          "author_name": "mrinath",
          "author_url": "",
          "post_date": "08/13/2021 17:55:15",
          "content": "<p>yet to use😅</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1471621,
      "author_name": "lillik",
      "author_url": "",
      "post_date": "08/14/2021 10:43:12",
      "content": "<p>There is a separate dataset for the segmentation part of the challenge (Task-1) here: <a href=\"https://www.synapse.org/#!Synapse:syn25829067/wiki/610863\" target=\"_blank\">https://www.synapse.org/#!Synapse:syn25829067/wiki/610863</a></p>\n<p>As far as I understand it, it‘s the same brains in nifti format, but with ground truth segmentation. I did not register for downloading it yet, but think it would be helpful. It would be ok to use that data for the classification part (Task-2) as well, right?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1464615": "What external datasets do you guys think may be relevant to this competition in terms of additional training data?\n\nThanks,",
    "1465163": "I found this one: [Brain Tumor Classification (MRI)](https://www.kaggle.com/sartajbhuvaji/brain-tumor-classification-mri)\nIt has 4 types of images, one type is no tumor, and another type is glioblastoma.\nBut the description says `you will predict the genetic subtype of glioblastoma using MRI (magnetic resonance imaging) scans`, I'm not sure if this glioma_tumor is as same as the glioblastoma tumor for this competition. \nThe external dataset combined T1, T2, FLAIR images as well, but you can try it.",
    "1465310": "Thanks for sharing!!! I think these data are good for training model",
    "1470601": "https://www.kaggle.com/adheshgarg/brats12\n\nthanks to https://www.kaggle.com/c/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/262120",
    "1470766": "I have viewed these images before. I thought they are too blurred to use. According to my experience, low-quality images sometimes damage the model's performance. Did this data improve your model?",
    "1470780": "yet to use😅",
    "1471621": "There is a separate dataset for the segmentation part of the challenge (Task-1) here: https://www.synapse.org/#!Synapse:syn25829067/wiki/610863\n\nAs far as I understand it, it‘s the same brains in nifti format, but with ground truth segmentation. I did not register for downloading it yet, but think it would be helpful. It would be ok to use that data for the classification part (Task-2) as well, right?"
  },
  "source": "meta"
}