{
  "id": 337257,
  "title": "Question about tiles",
  "url": "/competitions/hubmap-organ-segmentation/discussion/337257",
  "author_name": "",
  "post_date": "2022-07-15T07:26:36.672580Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Thanks for this <a href=\"url\" target=\"_blank\">https://www.kaggle.com/code/thedevastator/converting-to-256x256</a>, I learned the way that a large medical image can be divided into small tiles.<br>\nThis approach works well in some other competitions, but it doesn't work for me in this competitions.<br>\nI did spilt the large image into tiles and use them to train, but this way lead my model a negative effect, which is worse than I remain the large image.<br>\nI don't know if any one else has the same situation as me. Any advice is greatly appreciated.</p>",
  "messages": [
    {
      "id": "1856210",
      "postDate": "07/15/2022 07:26:36",
      "content": "<p>Thanks for this <a href=\"url\" target=\"_blank\">https://www.kaggle.com/code/thedevastator/converting-to-256x256</a>, I learned the way that a large medical image can be divided into small tiles.<br>\nThis approach works well in some other competitions, but it doesn't work for me in this competitions.<br>\nI did spilt the large image into tiles and use them to train, but this way lead my model a negative effect, which is worse than I remain the large image.<br>\nI don't know if any one else has the same situation as me. Any advice is greatly appreciated.</p>",
      "rawMarkdown": "Thanks for this [https://www.kaggle.com/code/thedevastator/converting-to-256x256](url), I learned the way that a large medical image can be divided into small tiles.\nThis approach works well in some other competitions, but it doesn't work for me in this competitions.\nI did spilt the large image into tiles and use them to train, but this way lead my model a negative effect, which is worse than I remain the large image.\nI don't know if any one else has the same situation as me. Any advice is greatly appreciated.",
      "votes": null
    },
    {
      "id": "1856292",
      "postDate": "07/15/2022 08:40:05",
      "content": "<p>Looking forward someone's relpy</p>",
      "rawMarkdown": "Looking forward someone's relpy",
      "votes": null
    },
    {
      "id": "1857386",
      "postDate": "07/16/2022 05:13:59",
      "content": "<p>I use RandomCrop for training and sliding window for inference. No great results so far</p>",
      "rawMarkdown": "I use RandomCrop for training and sliding window for inference. No great results so far",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1856292,
      "author_name": "jianxunzheng",
      "author_url": "",
      "post_date": "07/15/2022 08:40:05",
      "content": "<p>Looking forward someone's relpy</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1857386,
      "author_name": "quandapro",
      "author_url": "",
      "post_date": "07/16/2022 05:13:59",
      "content": "<p>I use RandomCrop for training and sliding window for inference. No great results so far</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1856210": "Thanks for this [https://www.kaggle.com/code/thedevastator/converting-to-256x256](url), I learned the way that a large medical image can be divided into small tiles.\nThis approach works well in some other competitions, but it doesn't work for me in this competitions.\nI did spilt the large image into tiles and use them to train, but this way lead my model a negative effect, which is worse than I remain the large image.\nI don't know if any one else has the same situation as me. Any advice is greatly appreciated.",
    "1856292": "Looking forward someone's relpy",
    "1857386": "I use RandomCrop for training and sliding window for inference. No great results so far"
  },
  "source": "meta"
}