{
  "id": 431439,
  "title": "Any Ideas on Segmentation?",
  "url": "/competitions/rsna-2023-abdominal-trauma-detection/discussion/431439",
  "author_name": "Jinyoung Seo",
  "post_date": "2023-08-13T16:31:43.498000",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello, I'm a newbie studying computer vision and I'm not familiar with using CT data, which is 3d…</p>\n<p>Is there any way to train the model without full image_level label? (Do you think using 3d model is more appropriate then using 2d convolution models?)<br>\nAlso, I was curious if there is a common way to use the segmentation data to make the model pay attention to the target organ for the corresponding label!<br>\nThank you!</p>",
  "messages": [
    {
      "id": 2388817,
      "postDate": "2023-08-13T16:31:43.500Z",
      "content": "<p>Hello, I'm a newbie studying computer vision and I'm not familiar with using CT data, which is 3d…</p>\n<p>Is there any way to train the model without full image_level label? (Do you think using 3d model is more appropriate then using 2d convolution models?)<br>\nAlso, I was curious if there is a common way to use the segmentation data to make the model pay attention to the target organ for the corresponding label!<br>\nThank you!</p>",
      "rawMarkdown": "Hello, I'm a newbie studying computer vision and I'm not familiar with using CT data, which is 3d...\n\nIs there any way to train the model without full image_level label? (Do you think using 3d model is more appropriate then using 2d convolution models?)\nAlso, I was curious if there is a common way to use the segmentation data to make the model pay attention to the target organ for the corresponding label!\nThank you!",
      "votes": 1
    },
    {
      "id": 2388933,
      "postDate": "2023-08-13T17:29:43.740Z",
      "content": "<ul>\n<li>I was wondering if the segmentation data really matters if there is only 206 data…and if the segmentation generated by total segment is reliable! (since there is no time to train both the classification and the segmentation model)</li>\n</ul>",
      "rawMarkdown": "+ I was wondering if the segmentation data really matters if there is only 206 data...and if the segmentation generated by total segment is reliable! (since there is no time to train both the classification and the segmentation model)"
    }
  ],
  "comments": [
    {
      "id": 2388933,
      "author_name": "Jinyoung Seo",
      "author_url": "",
      "post_date": "2023-08-13T17:29:43.740000",
      "content": "<ul>\n<li>I was wondering if the segmentation data really matters if there is only 206 data…and if the segmentation generated by total segment is reliable! (since there is no time to train both the classification and the segmentation model)</li>\n</ul>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2388817": "Hello, I'm a newbie studying computer vision and I'm not familiar with using CT data, which is 3d...\n\nIs there any way to train the model without full image_level label? (Do you think using 3d model is more appropriate then using 2d convolution models?)\nAlso, I was curious if there is a common way to use the segmentation data to make the model pay attention to the target organ for the corresponding label!\nThank you!",
    "2388933": "+ I was wondering if the segmentation data really matters if there is only 206 data...and if the segmentation generated by total segment is reliable! (since there is no time to train both the classification and the segmentation model)"
  }
}