{
  "id": 271787,
  "title": "Converting Between MRI Planes",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/271787",
  "author_name": "",
  "post_date": "2021-09-12T15:21:12.959850700Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hello everyone,</p>\n<p>I am assuming that the orientation that a mri scan takes when it is passed into a model should be constant. With that said, does anyone have advice on how I can take a scan in the coronal plane and transform it to either of the other 2 (or vice versa). Is it as simple as reinterpreting the 3d array or does some preprocessing need to be done in order to accomplish this task?</p>\n<p>Thanks,<br>\nSauman</p>",
  "messages": [
    {
      "id": "1510626",
      "postDate": "09/12/2021 15:21:12",
      "content": "<p>Hello everyone,</p>\n<p>I am assuming that the orientation that a mri scan takes when it is passed into a model should be constant. With that said, does anyone have advice on how I can take a scan in the coronal plane and transform it to either of the other 2 (or vice versa). Is it as simple as reinterpreting the 3d array or does some preprocessing need to be done in order to accomplish this task?</p>\n<p>Thanks,<br>\nSauman</p>",
      "rawMarkdown": "Hello everyone,\n\nI am assuming that the orientation that a mri scan takes when it is passed into a model should be constant. With that said, does anyone have advice on how I can take a scan in the coronal plane and transform it to either of the other 2 (or vice versa). Is it as simple as reinterpreting the 3d array or does some preprocessing need to be done in order to accomplish this task?\n\nThanks,\nSauman",
      "votes": null
    },
    {
      "id": "1511003",
      "postDate": "09/13/2021 02:42:42",
      "content": "<p>You can do it with SimpleITK. Take a look at this excellent notebook by <a href=\"https://www.kaggle.com/boojum\" target=\"_blank\">@boojum</a> -&gt; <a href=\"https://www.kaggle.com/boojum/connecting-voxel-spaces\" target=\"_blank\">https://www.kaggle.com/boojum/connecting-voxel-spaces</a></p>\n<p>Keep in mind that some series don't have many images, therefore the distance between slices is large. This results in poor quality reconstructed images.</p>",
      "rawMarkdown": "You can do it with SimpleITK. Take a look at this excellent notebook by @boojum -> https://www.kaggle.com/boojum/connecting-voxel-spaces\n\nKeep in mind that some series don't have many images, therefore the distance between slices is large. This results in poor quality reconstructed images.",
      "votes": null
    },
    {
      "id": "1511009",
      "postDate": "09/13/2021 02:52:54",
      "content": "<p>Thank you! I will definitely look into this notebook along with the limitations. </p>",
      "rawMarkdown": "Thank you! I will definitely look into this notebook along with the limitations.",
      "votes": null
    },
    {
      "id": "1513246",
      "postDate": "09/15/2021 02:25:44",
      "content": "<p>Thanks for asking <a href=\"https://www.kaggle.com/saumandas\" target=\"_blank\">@saumandas</a>! This really cleared my doubt too!</p>",
      "rawMarkdown": "Thanks for asking @saumandas! This really cleared my doubt too!",
      "votes": null
    },
    {
      "id": "1543621",
      "postDate": "10/13/2021 17:27:25",
      "content": "<p>how will your model improve if you decide to utilize mri planes (coronal, sagittal , axial) from dicom metadata , how would you even use in your model?</p>",
      "rawMarkdown": "how will your model improve if you decide to utilize mri planes (coronal, sagittal , axial) from dicom metadata , how would you even use in your model?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1511003,
      "author_name": "davidbroberts",
      "author_url": "",
      "post_date": "09/13/2021 02:42:42",
      "content": "<p>You can do it with SimpleITK. Take a look at this excellent notebook by <a href=\"https://www.kaggle.com/boojum\" target=\"_blank\">@boojum</a> -&gt; <a href=\"https://www.kaggle.com/boojum/connecting-voxel-spaces\" target=\"_blank\">https://www.kaggle.com/boojum/connecting-voxel-spaces</a></p>\n<p>Keep in mind that some series don't have many images, therefore the distance between slices is large. This results in poor quality reconstructed images.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1511009,
          "author_name": "saumandas",
          "author_url": "",
          "post_date": "09/13/2021 02:52:54",
          "content": "<p>Thank you! I will definitely look into this notebook along with the limitations. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1513246,
      "author_name": "conandoyle",
      "author_url": "",
      "post_date": "09/15/2021 02:25:44",
      "content": "<p>Thanks for asking <a href=\"https://www.kaggle.com/saumandas\" target=\"_blank\">@saumandas</a>! This really cleared my doubt too!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1543621,
      "author_name": "aniarya",
      "author_url": "",
      "post_date": "10/13/2021 17:27:25",
      "content": "<p>how will your model improve if you decide to utilize mri planes (coronal, sagittal , axial) from dicom metadata , how would you even use in your model?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1510626": "Hello everyone,\n\nI am assuming that the orientation that a mri scan takes when it is passed into a model should be constant. With that said, does anyone have advice on how I can take a scan in the coronal plane and transform it to either of the other 2 (or vice versa). Is it as simple as reinterpreting the 3d array or does some preprocessing need to be done in order to accomplish this task?\n\nThanks,\nSauman",
    "1511003": "You can do it with SimpleITK. Take a look at this excellent notebook by @boojum -> https://www.kaggle.com/boojum/connecting-voxel-spaces\n\nKeep in mind that some series don't have many images, therefore the distance between slices is large. This results in poor quality reconstructed images.",
    "1511009": "Thank you! I will definitely look into this notebook along with the limitations.",
    "1513246": "Thanks for asking @saumandas! This really cleared my doubt too!",
    "1543621": "how will your model improve if you decide to utilize mri planes (coronal, sagittal , axial) from dicom metadata , how would you even use in your model?"
  },
  "source": "meta"
}