{
  "id": 177675,
  "title": "How can I normalize the Hounsfield Units so that all of the CT scans are on the same scale?",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/177675",
  "author_name": "",
  "post_date": "2020-08-26T21:43:53.453476500Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>There seems to be a discrepancy in the Hounsfield Units. I tried to normalize the images based on the intercept, which some threads have suggested and by setting values of -2000 to 0, but these methods do not work for all of the images. Has anyone discovered a way to make sure that the CT images follow the Hounsfield Unit range of -1024 to 3071 HU?</p>",
  "messages": [
    {
      "id": "986911",
      "postDate": "08/26/2020 21:43:53",
      "content": "<p>There seems to be a discrepancy in the Hounsfield Units. I tried to normalize the images based on the intercept, which some threads have suggested and by setting values of -2000 to 0, but these methods do not work for all of the images. Has anyone discovered a way to make sure that the CT images follow the Hounsfield Unit range of -1024 to 3071 HU?</p>",
      "rawMarkdown": "There seems to be a discrepancy in the Hounsfield Units. I tried to normalize the images based on the intercept, which some threads have suggested and by setting values of -2000 to 0, but these methods do not work for all of the images. Has anyone discovered a way to make sure that the CT images follow the Hounsfield Unit range of -1024 to 3071 HU?",
      "votes": null
    },
    {
      "id": "987501",
      "postDate": "08/27/2020 10:02:58",
      "content": "<p>I'm seeing the same after rescaling by the files RescaleSlope and RescaleIntercept.<br>\nThis article:<br>\n<a href=\"https://radiopaedia.org/articles/hounsfield-unit?lang=us\" target=\"_blank\">https://radiopaedia.org/articles/hounsfield-unit?lang=us</a><br>\nspecifically mentions the -1024 to 3071 HU range only applies to 12-bit images. The dataset contains also 16-bit images, if I look at only the 12-bit scans I do get a consistent minimum value of -1024 but the maximum value is still distributed between +1000 and 3071. </p>",
      "rawMarkdown": "I'm seeing the same after rescaling by the files RescaleSlope and RescaleIntercept.\nThis article:\nhttps://radiopaedia.org/articles/hounsfield-unit?lang=us\nspecifically mentions the -1024 to 3071 HU range only applies to 12-bit images. The dataset contains also 16-bit images, if I look at only the 12-bit scans I do get a consistent minimum value of -1024 but the maximum value is still distributed between +1000 and 3071.",
      "votes": null
    },
    {
      "id": "1013637",
      "postDate": "09/16/2020 20:45:27",
      "content": "<p>What is the significance between the 16-bit and 12-bit scans? That is an interesting discovery</p>",
      "rawMarkdown": "What is the significance between the 16-bit and 12-bit scans? That is an interesting discovery",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 987501,
      "author_name": "bart01",
      "author_url": "",
      "post_date": "08/27/2020 10:02:58",
      "content": "<p>I'm seeing the same after rescaling by the files RescaleSlope and RescaleIntercept.<br>\nThis article:<br>\n<a href=\"https://radiopaedia.org/articles/hounsfield-unit?lang=us\" target=\"_blank\">https://radiopaedia.org/articles/hounsfield-unit?lang=us</a><br>\nspecifically mentions the -1024 to 3071 HU range only applies to 12-bit images. The dataset contains also 16-bit images, if I look at only the 12-bit scans I do get a consistent minimum value of -1024 but the maximum value is still distributed between +1000 and 3071. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1013637,
          "author_name": "akiroduey",
          "author_url": "",
          "post_date": "09/16/2020 20:45:27",
          "content": "<p>What is the significance between the 16-bit and 12-bit scans? That is an interesting discovery</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "986911": "There seems to be a discrepancy in the Hounsfield Units. I tried to normalize the images based on the intercept, which some threads have suggested and by setting values of -2000 to 0, but these methods do not work for all of the images. Has anyone discovered a way to make sure that the CT images follow the Hounsfield Unit range of -1024 to 3071 HU?",
    "987501": "I'm seeing the same after rescaling by the files RescaleSlope and RescaleIntercept.\nThis article:\nhttps://radiopaedia.org/articles/hounsfield-unit?lang=us\nspecifically mentions the -1024 to 3071 HU range only applies to 12-bit images. The dataset contains also 16-bit images, if I look at only the 12-bit scans I do get a consistent minimum value of -1024 but the maximum value is still distributed between +1000 and 3071.",
    "1013637": "What is the significance between the 16-bit and 12-bit scans? That is an interesting discovery"
  },
  "source": "meta"
}