{
  "id": 271572,
  "title": "What problems does this dataset have?",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/271572",
  "author_name": "alckasoc",
  "post_date": "2021-09-11T08:34:48.597000",
  "votes": 7,
  "comment_count": 0,
  "views": 0,
  "content": "<p>My team and I have come across a few general problems for this dataset.<br>\nSome including:</p>\n<ul>\n<li>small test dataset</li>\n<li>small public dataset</li>\n<li>black images</li>\n<li>patient folders that need to be ignored</li>\n<li>image scans that are too bright to be used</li>\n<li>maybe angling issues with the scans?</li>\n<li>too much of the black background </li>\n<li>don't know which slice of the image stack is most indicative of the methylation</li>\n<li>no clear cut idea of what information in the images for each mpMRI scan type to enhance to bring out the most information (e.g. if one scan type was for one specific type of information like water then how can we enhance that)</li>\n<li>there may be more problems hidden in the train set…</li>\n</ul>\n<p>Feel free to add to the list or correct me! </p>",
  "messages": [
    {
      "id": 1509397,
      "postDate": "2021-09-11T08:34:48.597Z",
      "content": "<p>My team and I have come across a few general problems for this dataset.<br>\nSome including:</p>\n<ul>\n<li>small test dataset</li>\n<li>small public dataset</li>\n<li>black images</li>\n<li>patient folders that need to be ignored</li>\n<li>image scans that are too bright to be used</li>\n<li>maybe angling issues with the scans?</li>\n<li>too much of the black background </li>\n<li>don't know which slice of the image stack is most indicative of the methylation</li>\n<li>no clear cut idea of what information in the images for each mpMRI scan type to enhance to bring out the most information (e.g. if one scan type was for one specific type of information like water then how can we enhance that)</li>\n<li>there may be more problems hidden in the train set…</li>\n</ul>\n<p>Feel free to add to the list or correct me! </p>",
      "rawMarkdown": "My team and I have come across a few general problems for this dataset.\nSome including:\n- small test dataset\n- small public dataset\n- black images\n- patient folders that need to be ignored\n- image scans that are too bright to be used\n- maybe angling issues with the scans?\n- too much of the black background \n- don't know which slice of the image stack is most indicative of the methylation\n- no clear cut idea of what information in the images for each mpMRI scan type to enhance to bring out the most information (e.g. if one scan type was for one specific type of information like water then how can we enhance that)\n- there may be more problems hidden in the train set...\n\nFeel free to add to the list or correct me! ",
      "votes": 7
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1509397": "My team and I have come across a few general problems for this dataset.\nSome including:\n- small test dataset\n- small public dataset\n- black images\n- patient folders that need to be ignored\n- image scans that are too bright to be used\n- maybe angling issues with the scans?\n- too much of the black background \n- don't know which slice of the image stack is most indicative of the methylation\n- no clear cut idea of what information in the images for each mpMRI scan type to enhance to bring out the most information (e.g. if one scan type was for one specific type of information like water then how can we enhance that)\n- there may be more problems hidden in the train set...\n\nFeel free to add to the list or correct me! "
  }
}