{
  "id": 244808,
  "title": "bounding box- can somebody explain EACH number?  ",
  "url": "/competitions/siim-covid19-detection/discussion/244808",
  "author_name": "",
  "post_date": "2021-06-08T11:23:09.420187300Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I think two are latitude, two are longitude, and one is a confidence interval?</p>\n<p>is this written somewhere, or is it assumed that we all know this?</p>\n<p>thanks very much</p>",
  "messages": [
    {
      "id": "1341029",
      "postDate": "06/08/2021 11:23:09",
      "content": "<p>I think two are latitude, two are longitude, and one is a confidence interval?</p>\n<p>is this written somewhere, or is it assumed that we all know this?</p>\n<p>thanks very much</p>",
      "rawMarkdown": "I think two are latitude, two are longitude, and one is a confidence interval?\n\nis this written somewhere, or is it assumed that we all know this?\n\nthanks very much",
      "votes": null
    },
    {
      "id": "1341331",
      "postDate": "06/08/2021 15:33:10",
      "content": "<p>For the train data ..</p>\n<p>The variables in the boxes column are x, y, width, height and label .. where x and y are the pixel coordinates of the top left corner.</p>\n<p>There can be more than one box per image, so these will be in an array of dicts.</p>\n<p>For predictions,  here's what the data page says about it.</p>\n<blockquote>\n  <p>If you predict that there are no findings, you should create a prediction of \"none 1 0 0 1 1\" (\"none\" is the class ID for no finding, and this provides a one-pixel bounding box with a confidence of 1.0).</p>\n</blockquote>\n<p><a href=\"https://www.kaggle.com/c/siim-covid19-detection/data\" target=\"_blank\">https://www.kaggle.com/c/siim-covid19-detection/data</a></p>",
      "rawMarkdown": "For the train data ..\n\nThe variables in the boxes column are x, y, width, height and label .. where x and y are the pixel coordinates of the top left corner.\n\nThere can be more than one box per image, so these will be in an array of dicts.\n\nFor predictions,  here's what the data page says about it.\n\n> If you predict that there are no findings, you should create a prediction of \"none 1 0 0 1 1\" (\"none\" is the class ID for no finding, and this provides a one-pixel bounding box with a confidence of 1.0).\n\nhttps://www.kaggle.com/c/siim-covid19-detection/data",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1341331,
      "author_name": "davidbroberts",
      "author_url": "",
      "post_date": "06/08/2021 15:33:10",
      "content": "<p>For the train data ..</p>\n<p>The variables in the boxes column are x, y, width, height and label .. where x and y are the pixel coordinates of the top left corner.</p>\n<p>There can be more than one box per image, so these will be in an array of dicts.</p>\n<p>For predictions,  here's what the data page says about it.</p>\n<blockquote>\n  <p>If you predict that there are no findings, you should create a prediction of \"none 1 0 0 1 1\" (\"none\" is the class ID for no finding, and this provides a one-pixel bounding box with a confidence of 1.0).</p>\n</blockquote>\n<p><a href=\"https://www.kaggle.com/c/siim-covid19-detection/data\" target=\"_blank\">https://www.kaggle.com/c/siim-covid19-detection/data</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1341029": "I think two are latitude, two are longitude, and one is a confidence interval?\n\nis this written somewhere, or is it assumed that we all know this?\n\nthanks very much",
    "1341331": "For the train data ..\n\nThe variables in the boxes column are x, y, width, height and label .. where x and y are the pixel coordinates of the top left corner.\n\nThere can be more than one box per image, so these will be in an array of dicts.\n\nFor predictions,  here's what the data page says about it.\n\n> If you predict that there are no findings, you should create a prediction of \"none 1 0 0 1 1\" (\"none\" is the class ID for no finding, and this provides a one-pixel bounding box with a confidence of 1.0).\n\nhttps://www.kaggle.com/c/siim-covid19-detection/data"
  },
  "source": "meta"
}