{
  "id": 438194,
  "title": "[Beginner] How to approach this problem?",
  "url": "/competitions/rsna-2023-abdominal-trauma-detection/discussion/438194",
  "author_name": "Eddwait",
  "post_date": "2023-09-09T23:53:53.783000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello everyone, I have been trying to figure out a way to go about this problem. So, I stumbled upon 3d images and learned that their voxel values ranges from -1000 to 400 and above. However, here after performing eda I found out that the values range from anywhere between -8000 (approx) to 30000+ (approx.) and this goes for the lower-bound voxel values. Meaning, the values that were supposed to be -1000 are 30000 and above! And it's different for different scenarios (for the most it's -1024, -2048, 0). So, here's my question, how do I normalize these images? What should be the basic step after that (going about building the model). Also, the segmentation volumes have been given if any of you are able to make sense of that, how are you guys leveraging it? Thank you in advance!</p>",
  "messages": [
    {
      "id": 2431243,
      "postDate": "2023-09-09T23:53:53.783Z",
      "content": "<p>Hello everyone, I have been trying to figure out a way to go about this problem. So, I stumbled upon 3d images and learned that their voxel values ranges from -1000 to 400 and above. However, here after performing eda I found out that the values range from anywhere between -8000 (approx) to 30000+ (approx.) and this goes for the lower-bound voxel values. Meaning, the values that were supposed to be -1000 are 30000 and above! And it's different for different scenarios (for the most it's -1024, -2048, 0). So, here's my question, how do I normalize these images? What should be the basic step after that (going about building the model). Also, the segmentation volumes have been given if any of you are able to make sense of that, how are you guys leveraging it? Thank you in advance!</p>",
      "rawMarkdown": "Hello everyone, I have been trying to figure out a way to go about this problem. So, I stumbled upon 3d images and learned that their voxel values ranges from -1000 to 400 and above. However, here after performing eda I found out that the values range from anywhere between -8000 (approx) to 30000+ (approx.) and this goes for the lower-bound voxel values. Meaning, the values that were supposed to be -1000 are 30000 and above! And it's different for different scenarios (for the most it's -1024, -2048, 0). So, here's my question, how do I normalize these images? What should be the basic step after that (going about building the model). Also, the segmentation volumes have been given if any of you are able to make sense of that, how are you guys leveraging it? Thank you in advance!"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2431243": "Hello everyone, I have been trying to figure out a way to go about this problem. So, I stumbled upon 3d images and learned that their voxel values ranges from -1000 to 400 and above. However, here after performing eda I found out that the values range from anywhere between -8000 (approx) to 30000+ (approx.) and this goes for the lower-bound voxel values. Meaning, the values that were supposed to be -1000 are 30000 and above! And it's different for different scenarios (for the most it's -1024, -2048, 0). So, here's my question, how do I normalize these images? What should be the basic step after that (going about building the model). Also, the segmentation volumes have been given if any of you are able to make sense of that, how are you guys leveraging it? Thank you in advance!"
  }
}