{
  "id": 188737,
  "title": "How to choice image for training?",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/188737",
  "author_name": "",
  "post_date": "2020-10-05T02:58:15.647679100Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>The image below of a patient, The image 1,2,3,22,23,24,25,26,27,28,29,30 doesn't matter much information. And It makes noise data for training. What is the better way for training data?</p>\n<p>Reference: <a href=\"https://www.kaggle.com/andradaolteanu/pulmonary-fibrosis-competition-eda-dicom-prep\" target=\"_blank\">https://www.kaggle.com/andradaolteanu/pulmonary-fibrosis-competition-eda-dicom-prep</a><br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4000197%2F2844597d2704f5992a013d5dfa8f0f9e%2Fa.png?generation=1601867016508865&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": "1037436",
      "postDate": "10/05/2020 02:58:15",
      "content": "<p>The image below of a patient, The image 1,2,3,22,23,24,25,26,27,28,29,30 doesn't matter much information. And It makes noise data for training. What is the better way for training data?</p>\n<p>Reference: <a href=\"https://www.kaggle.com/andradaolteanu/pulmonary-fibrosis-competition-eda-dicom-prep\" target=\"_blank\">https://www.kaggle.com/andradaolteanu/pulmonary-fibrosis-competition-eda-dicom-prep</a><br>\n<img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4000197%2F2844597d2704f5992a013d5dfa8f0f9e%2Fa.png?generation=1601867016508865&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "The image below of a patient, The image 1,2,3,22,23,24,25,26,27,28,29,30 doesn't matter much information. And It makes noise data for training. What is the better way for training data?\n\nReference: https://www.kaggle.com/andradaolteanu/pulmonary-fibrosis-competition-eda-dicom-prep\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4000197%2F2844597d2704f5992a013d5dfa8f0f9e%2Fa.png?generation=1601867016508865&alt=media)",
      "votes": null
    },
    {
      "id": "1039113",
      "postDate": "10/06/2020 10:42:04",
      "content": "<p>You have 2 options if you wnat to deal witgh images:</p>\n<ul>\n<li>Build 2d image model</li>\n<li>Build 3d image model</li>\n</ul>\n<p>In case you choose 2d image model you will have to face the problem about how to choose the most relevant image or maybe to loop over images and append latent features to sparse vector…</p>\n<p>On my personal opinion 3d image model is more natural, even it is heavier.</p>\n<p>Good luck!</p>",
      "rawMarkdown": "You have 2 options if you wnat to deal witgh images:\n\n- Build 2d image model\n- Build 3d image model\n\nIn case you choose 2d image model you will have to face the problem about how to choose the most relevant image or maybe to loop over images and append latent features to sparse vector...\n\nOn my personal opinion 3d image model is more natural, even it is heavier.\n\nGood luck!",
      "votes": null
    },
    {
      "id": "1040175",
      "postDate": "10/07/2020 02:42:17",
      "content": "<p><a href=\"https://www.kaggle.com/enric1296\" target=\"_blank\">@enric1296</a>  thank you 👍👍👍</p>",
      "rawMarkdown": "enric1296  thank you 👍👍👍",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1039113,
      "author_name": "enric1296",
      "author_url": "",
      "post_date": "10/06/2020 10:42:04",
      "content": "<p>You have 2 options if you wnat to deal witgh images:</p>\n<ul>\n<li>Build 2d image model</li>\n<li>Build 3d image model</li>\n</ul>\n<p>In case you choose 2d image model you will have to face the problem about how to choose the most relevant image or maybe to loop over images and append latent features to sparse vector…</p>\n<p>On my personal opinion 3d image model is more natural, even it is heavier.</p>\n<p>Good luck!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1040175,
      "author_name": "sentrankim",
      "author_url": "",
      "post_date": "10/07/2020 02:42:17",
      "content": "<p><a href=\"https://www.kaggle.com/enric1296\" target=\"_blank\">@enric1296</a>  thank you 👍👍👍</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1037436": "The image below of a patient, The image 1,2,3,22,23,24,25,26,27,28,29,30 doesn't matter much information. And It makes noise data for training. What is the better way for training data?\n\nReference: https://www.kaggle.com/andradaolteanu/pulmonary-fibrosis-competition-eda-dicom-prep\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4000197%2F2844597d2704f5992a013d5dfa8f0f9e%2Fa.png?generation=1601867016508865&alt=media)",
    "1039113": "You have 2 options if you wnat to deal witgh images:\n\n- Build 2d image model\n- Build 3d image model\n\nIn case you choose 2d image model you will have to face the problem about how to choose the most relevant image or maybe to loop over images and append latent features to sparse vector...\n\nOn my personal opinion 3d image model is more natural, even it is heavier.\n\nGood luck!",
    "1040175": "enric1296  thank you 👍👍👍"
  },
  "source": "meta"
}