{
  "id": 53523,
  "title": "submission rows for the same imageId",
  "url": "/competitions/data-science-bowl-2018/discussion/53523",
  "author_name": "AyakZob",
  "post_date": "2018-04-01T02:30:24.088000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>It is not very clear to me how each row is calculated in the submission file. I am planning to use u-net to segment the image. I am going to merge the masks and train the network to find a single output mask. From there what is the logic of creating each row for the submission file. Is it one row for each contiguous region or is it one row for each nucleus even if multiple nuclei are found in one contiguous region? </p>",
  "messages": [
    {
      "id": 307173,
      "postDate": "2018-04-01T02:30:24.090Z",
      "content": "<p>It is not very clear to me how each row is calculated in the submission file. I am planning to use u-net to segment the image. I am going to merge the masks and train the network to find a single output mask. From there what is the logic of creating each row for the submission file. Is it one row for each contiguous region or is it one row for each nucleus even if multiple nuclei are found in one contiguous region? </p>",
      "rawMarkdown": "It is not very clear to me how each row is calculated in the submission file. I am planning to use u-net to segment the image. I am going to merge the masks and train the network to find a single output mask. From there what is the logic of creating each row for the submission file. Is it one row for each contiguous region or is it one row for each nucleus even if multiple nuclei are found in one contiguous region? "
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "307173": "It is not very clear to me how each row is calculated in the submission file. I am planning to use u-net to segment the image. I am going to merge the masks and train the network to find a single output mask. From there what is the logic of creating each row for the submission file. Is it one row for each contiguous region or is it one row for each nucleus even if multiple nuclei are found in one contiguous region? "
  }
}