{
  "id": 252931,
  "title": "What is `BraTS21ID`in training CSV and how does it map to the actual Images?",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/252931",
  "author_name": "",
  "post_date": "2021-07-14T09:16:30.979632Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>In the training dataset, there is a given column <code>BraTS21ID</code> with the training CSV, which I cannot understand as to where the actual Image lies.</p>\n<p>The understanding of the directory structure that I have is :</p>\n<h3>Directory Structure</h3>\n<pre><code>Test/Train Directory\n    Study_ID_Number - 00001, ....\n        FLAIR\n        T1w\n        T1wCE\n        T2w\n            Image-{Img_ID}.dcm\n</code></pre>\n<p>For example an ID of <code>BraTS21ID</code> set as 2 would point to which Image in the Training Dataset?</p>",
  "messages": [
    {
      "id": "1387565",
      "postDate": "07/14/2021 09:16:30",
      "content": "<p>In the training dataset, there is a given column <code>BraTS21ID</code> with the training CSV, which I cannot understand as to where the actual Image lies.</p>\n<p>The understanding of the directory structure that I have is :</p>\n<h3>Directory Structure</h3>\n<pre><code>Test/Train Directory\n    Study_ID_Number - 00001, ....\n        FLAIR\n        T1w\n        T1wCE\n        T2w\n            Image-{Img_ID}.dcm\n</code></pre>\n<p>For example an ID of <code>BraTS21ID</code> set as 2 would point to which Image in the Training Dataset?</p>",
      "rawMarkdown": "In the training dataset, there is a given column `BraTS21ID` with the training CSV, which I cannot understand as to where the actual Image lies.\n\nThe understanding of the directory structure that I have is :\n\n###  Directory Structure\n\n```\nTest/Train Directory\n    Study_ID_Number - 00001, ....\n        FLAIR\n        T1w\n        T1wCE\n        T2w\n            Image-{Img_ID}.dcm\n```\n\nFor example an ID of `BraTS21ID` set as 2 would point to which Image in the Training Dataset?",
      "votes": null
    },
    {
      "id": "1387568",
      "postDate": "07/14/2021 09:20:19",
      "content": "<p>So I get to understand this like this : Each Study ID has the same Class. This is the only Plausible Explanation of the Data.</p>",
      "rawMarkdown": "So I get to understand this like this : Each Study ID has the same Class. This is the only Plausible Explanation of the Data.",
      "votes": null
    },
    {
      "id": "1387833",
      "postDate": "07/14/2021 13:02:27",
      "content": "<p>I think BraTS21ID is short for \"Brain Tumor Segmentation 2021 ID\" .. which is the id for each study.</p>\n<blockquote>\n  <p>For example an ID of BraTS21ID set as 2 would point to which Image in the Training Dataset?</p>\n</blockquote>\n<p>The ID doesn't point to an image, rather a study.  2 would actually look like this (need to add leading zeros) ..</p>\n<p><code>../input/rsna-miccai-brain-tumor-radiogenomic-classification/train/00002</code></p>\n<p>Inside that directory are the four series directories (FLAIR, T1w, T1wCE, T2w).</p>",
      "rawMarkdown": "I think BraTS21ID is short for \"Brain Tumor Segmentation 2021 ID\" .. which is the id for each study.\n\n> For example an ID of BraTS21ID set as 2 would point to which Image in the Training Dataset?\n\nThe ID doesn't point to an image, rather a study.  2 would actually look like this (need to add leading zeros) ..\n\n`../input/rsna-miccai-brain-tumor-radiogenomic-classification/train/00002`\n\nInside that directory are the four series directories (FLAIR, T1w, T1wCE, T2w).",
      "votes": null
    },
    {
      "id": "1453887",
      "postDate": "08/06/2021 03:20:29",
      "content": "<p>I was trying to understand the same thing and I read your coment and your right. The only thing I still don´t understand is what is there in each \"study\".  With the same example of study 00002, we have 4 sub directories (FLAIR, T1w, T1wCE, T2w) in which there are uneven amounts of images. Specifically in study 2 we have: 129 in FLAIR, 31 in T1w, 129 in T1wCE and 384 in T2w. So my question is, each study corresponds to a particular person, and that person has all those images related to him/her?</p>",
      "rawMarkdown": "I was trying to understand the same thing and I read your coment and your right. The only thing I still don´t understand is what is there in each \"study\".  With the same example of study 00002, we have 4 sub directories (FLAIR, T1w, T1wCE, T2w) in which there are uneven amounts of images. Specifically in study 2 we have: 129 in FLAIR, 31 in T1w, 129 in T1wCE and 384 in T2w. So my question is, each study corresponds to a particular person, and that person has all those images related to him/her?",
      "votes": null
    },
    {
      "id": "1453961",
      "postDate": "08/06/2021 03:59:29",
      "content": "<p><a href=\"https://www.kaggle.com/wernerignaciogonzlez\" target=\"_blank\">@wernerignaciogonzlez</a>, that is correct. Each study contains four series on the same patient during the same 'exam'. Each series in these studies represents a different acquisition designed to demonstrate the contrast between anatomical structure densities slightly differently. Tissues have different relative signal intensities between series. </p>\n<p>There are various technical reasons for different numbers of images in each series. Mostly its because these images were likely protocoled for a human to read and it isn't always diagnostically necessary to view each series in fine detail.</p>",
      "rawMarkdown": "wernerignaciogonzlez, that is correct. Each study contains four series on the same patient during the same 'exam'. Each series in these studies represents a different acquisition designed to demonstrate the contrast between anatomical structure densities slightly differently. Tissues have different relative signal intensities between series. \n\nThere are various technical reasons for different numbers of images in each series. Mostly its because these images were likely protocoled for a human to read and it isn't always diagnostically necessary to view each series in fine detail.",
      "votes": null
    },
    {
      "id": "1454546",
      "postDate": "08/06/2021 08:33:01",
      "content": "<p>Thanks David! Needed to verify that :D</p>",
      "rawMarkdown": "Thanks David! Needed to verify that :D",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1387568,
      "author_name": "farhanhaikhan",
      "author_url": "",
      "post_date": "07/14/2021 09:20:19",
      "content": "<p>So I get to understand this like this : Each Study ID has the same Class. This is the only Plausible Explanation of the Data.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1387833,
      "author_name": "davidbroberts",
      "author_url": "",
      "post_date": "07/14/2021 13:02:27",
      "content": "<p>I think BraTS21ID is short for \"Brain Tumor Segmentation 2021 ID\" .. which is the id for each study.</p>\n<blockquote>\n  <p>For example an ID of BraTS21ID set as 2 would point to which Image in the Training Dataset?</p>\n</blockquote>\n<p>The ID doesn't point to an image, rather a study.  2 would actually look like this (need to add leading zeros) ..</p>\n<p><code>../input/rsna-miccai-brain-tumor-radiogenomic-classification/train/00002</code></p>\n<p>Inside that directory are the four series directories (FLAIR, T1w, T1wCE, T2w).</p>",
      "votes": null,
      "replies": [
        {
          "id": 1453887,
          "author_name": "wernerignaciogonzlez",
          "author_url": "",
          "post_date": "08/06/2021 03:20:29",
          "content": "<p>I was trying to understand the same thing and I read your coment and your right. The only thing I still don´t understand is what is there in each \"study\".  With the same example of study 00002, we have 4 sub directories (FLAIR, T1w, T1wCE, T2w) in which there are uneven amounts of images. Specifically in study 2 we have: 129 in FLAIR, 31 in T1w, 129 in T1wCE and 384 in T2w. So my question is, each study corresponds to a particular person, and that person has all those images related to him/her?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1453961,
          "author_name": "davidbroberts",
          "author_url": "",
          "post_date": "08/06/2021 03:59:29",
          "content": "<p><a href=\"https://www.kaggle.com/wernerignaciogonzlez\" target=\"_blank\">@wernerignaciogonzlez</a>, that is correct. Each study contains four series on the same patient during the same 'exam'. Each series in these studies represents a different acquisition designed to demonstrate the contrast between anatomical structure densities slightly differently. Tissues have different relative signal intensities between series. </p>\n<p>There are various technical reasons for different numbers of images in each series. Mostly its because these images were likely protocoled for a human to read and it isn't always diagnostically necessary to view each series in fine detail.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1454546,
          "author_name": "wernerignaciogonzlez",
          "author_url": "",
          "post_date": "08/06/2021 08:33:01",
          "content": "<p>Thanks David! Needed to verify that :D</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1387565": "In the training dataset, there is a given column `BraTS21ID` with the training CSV, which I cannot understand as to where the actual Image lies.\n\nThe understanding of the directory structure that I have is :\n\n###  Directory Structure\n\n```\nTest/Train Directory\n    Study_ID_Number - 00001, ....\n        FLAIR\n        T1w\n        T1wCE\n        T2w\n            Image-{Img_ID}.dcm\n```\n\nFor example an ID of `BraTS21ID` set as 2 would point to which Image in the Training Dataset?",
    "1387568": "So I get to understand this like this : Each Study ID has the same Class. This is the only Plausible Explanation of the Data.",
    "1387833": "I think BraTS21ID is short for \"Brain Tumor Segmentation 2021 ID\" .. which is the id for each study.\n\n> For example an ID of BraTS21ID set as 2 would point to which Image in the Training Dataset?\n\nThe ID doesn't point to an image, rather a study.  2 would actually look like this (need to add leading zeros) ..\n\n`../input/rsna-miccai-brain-tumor-radiogenomic-classification/train/00002`\n\nInside that directory are the four series directories (FLAIR, T1w, T1wCE, T2w).",
    "1453887": "I was trying to understand the same thing and I read your coment and your right. The only thing I still don´t understand is what is there in each \"study\".  With the same example of study 00002, we have 4 sub directories (FLAIR, T1w, T1wCE, T2w) in which there are uneven amounts of images. Specifically in study 2 we have: 129 in FLAIR, 31 in T1w, 129 in T1wCE and 384 in T2w. So my question is, each study corresponds to a particular person, and that person has all those images related to him/her?",
    "1453961": "wernerignaciogonzlez, that is correct. Each study contains four series on the same patient during the same 'exam'. Each series in these studies represents a different acquisition designed to demonstrate the contrast between anatomical structure densities slightly differently. Tissues have different relative signal intensities between series. \n\nThere are various technical reasons for different numbers of images in each series. Mostly its because these images were likely protocoled for a human to read and it isn't always diagnostically necessary to view each series in fine detail.",
    "1454546": "Thanks David! Needed to verify that :D"
  },
  "source": "meta"
}