{
  "id": 213523,
  "title": "Paper Vs Competition Labels",
  "url": "/competitions/vinbigdata-chest-xray-abnormalities-detection/discussion/213523",
  "author_name": "Awsaf",
  "post_date": "2021-01-23T08:16:39.177000",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>According to the paper there are <code>22</code> <strong>local</strong> and <code>6</code> <strong>global labels</strong> but for the competition we have only <code>14</code>(without <strong>No Finding</strong>). I couldn't find any explanation how <code>22</code> labels were converted to <code>14</code>. Count of only <strong>local labels</strong> doesn't match with the count of the <strong>competition labels</strong>. <br>\n<a href=\"https://www.kaggle.com/juliaelliott\" target=\"_blank\">@juliaelliott</a> <a href=\"https://www.kaggle.com/nguyenquyha\" target=\"_blank\">@nguyenquyha</a> <a href=\"https://www.kaggle.com/nguyenbadung\" target=\"_blank\">@nguyenbadung</a> could  you provide some explanation?</p>",
  "messages": [
    {
      "id": 1165788,
      "postDate": "2021-01-23T08:32:23.980Z",
      "content": "<p>Since there are very few samples annotated with the other 8 classes, those labels were merged into the class \"<strong>other lesion</strong>\".</p>",
      "rawMarkdown": "Since there are very few samples annotated with the other 8 classes, those labels were merged into the class \"**other lesion**\".",
      "votes": 6
    },
    {
      "id": 1165766,
      "postDate": "2021-01-23T08:16:39.177Z",
      "content": "<p>According to the paper there are <code>22</code> <strong>local</strong> and <code>6</code> <strong>global labels</strong> but for the competition we have only <code>14</code>(without <strong>No Finding</strong>). I couldn't find any explanation how <code>22</code> labels were converted to <code>14</code>. Count of only <strong>local labels</strong> doesn't match with the count of the <strong>competition labels</strong>. <br>\n<a href=\"https://www.kaggle.com/juliaelliott\" target=\"_blank\">@juliaelliott</a> <a href=\"https://www.kaggle.com/nguyenquyha\" target=\"_blank\">@nguyenquyha</a> <a href=\"https://www.kaggle.com/nguyenbadung\" target=\"_blank\">@nguyenbadung</a> could  you provide some explanation?</p>",
      "rawMarkdown": "According to the paper there are `22` **local** and `6` **global labels** but for the competition we have only `14`(without **No Finding**). I couldn't find any explanation how `22` labels were converted to `14`. Count of only **local labels** doesn't match with the count of the **competition labels**. \n@juliaelliott @nguyenquyha @nguyenbadung could  you provide some explanation?",
      "votes": 3
    },
    {
      "id": 1168945,
      "postDate": "2021-01-25T09:01:10.830Z",
      "content": "<p>5 radiologists label the test set were selected from the pool of 17 radiologists? <a href=\"https://www.kaggle.com/nguyenquyha\" target=\"_blank\">@nguyenquyha</a> </p>",
      "rawMarkdown": "5 radiologists label the test set were selected from the pool of 17 radiologists? @nguyenquyha "
    }
  ],
  "comments": [
    {
      "id": 1165788,
      "author_name": "Ha Q. Nguyen",
      "author_url": "",
      "post_date": "2021-01-23T08:32:23.980000",
      "content": "<p>Since there are very few samples annotated with the other 8 classes, those labels were merged into the class \"<strong>other lesion</strong>\".</p>",
      "votes": 6,
      "replies": []
    },
    {
      "id": 1168945,
      "author_name": "Nguyen Phuc Dat",
      "author_url": "",
      "post_date": "2021-01-25T09:01:10.830000",
      "content": "<p>5 radiologists label the test set were selected from the pool of 17 radiologists? <a href=\"https://www.kaggle.com/nguyenquyha\" target=\"_blank\">@nguyenquyha</a> </p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1165788": "Since there are very few samples annotated with the other 8 classes, those labels were merged into the class \"**other lesion**\".",
    "1165766": "According to the paper there are `22` **local** and `6` **global labels** but for the competition we have only `14`(without **No Finding**). I couldn't find any explanation how `22` labels were converted to `14`. Count of only **local labels** doesn't match with the count of the **competition labels**. \n@juliaelliott @nguyenquyha @nguyenbadung could  you provide some explanation?",
    "1168945": "5 radiologists label the test set were selected from the pool of 17 radiologists? @nguyenquyha "
  }
}