{
  "id": 353264,
  "title": "How to use train annotations",
  "url": "/competitions/hubmap-organ-segmentation/discussion/353264",
  "author_name": "",
  "post_date": "2022-09-17T14:55:46.176258400Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I arrive late in this competition, I wonder how we use train annotations?</p>",
  "messages": [
    {
      "id": "1943495",
      "postDate": "09/17/2022 14:55:46",
      "content": "<p>I arrive late in this competition, I wonder how we use train annotations?</p>",
      "rawMarkdown": "I arrive late in this competition, I wonder how we use train annotations?",
      "votes": null
    },
    {
      "id": "1943514",
      "postDate": "09/17/2022 15:07:07",
      "content": "<p>excerpt from <a href=\"https://www.kaggle.com/competitions/hubmap-organ-segmentation/discussion/332958#1833181\" target=\"_blank\">Post</a>, <code>The json versions are more raw. The annotations might have issues that don't exist in the RLE copies, like overlaps, but can also allow you to distinguish between multiple adjacent FTUs which would all end up in the same mask with RLE.</code>. So I don't think anyone use train annotations. I could be wrong though.</p>\n<p>PS the hightlight link doesn't properly go to Mr. Sohier Dane quotes. Sorry for the inconvenience.</p>",
      "rawMarkdown": "excerpt from [Post](https://www.kaggle.com/competitions/hubmap-organ-segmentation/discussion/332958#1833181), ```The json versions are more raw. The annotations might have issues that don't exist in the RLE copies, like overlaps, but can also allow you to distinguish between multiple adjacent FTUs which would all end up in the same mask with RLE.```. So I don't think anyone use train annotations. I could be wrong though.\n\nPS the hightlight link doesn't properly go to Mr. Sohier Dane quotes. Sorry for the inconvenience.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1943514,
      "author_name": "cheulkay",
      "author_url": "",
      "post_date": "09/17/2022 15:07:07",
      "content": "<p>excerpt from <a href=\"https://www.kaggle.com/competitions/hubmap-organ-segmentation/discussion/332958#1833181\" target=\"_blank\">Post</a>, <code>The json versions are more raw. The annotations might have issues that don't exist in the RLE copies, like overlaps, but can also allow you to distinguish between multiple adjacent FTUs which would all end up in the same mask with RLE.</code>. So I don't think anyone use train annotations. I could be wrong though.</p>\n<p>PS the hightlight link doesn't properly go to Mr. Sohier Dane quotes. Sorry for the inconvenience.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1943495": "I arrive late in this competition, I wonder how we use train annotations?",
    "1943514": "excerpt from [Post](https://www.kaggle.com/competitions/hubmap-organ-segmentation/discussion/332958#1833181), ```The json versions are more raw. The annotations might have issues that don't exist in the RLE copies, like overlaps, but can also allow you to distinguish between multiple adjacent FTUs which would all end up in the same mask with RLE.```. So I don't think anyone use train annotations. I could be wrong though.\n\nPS the hightlight link doesn't properly go to Mr. Sohier Dane quotes. Sorry for the inconvenience."
  },
  "source": "meta"
}