{
  "id": 63604,
  "title": "How to convert to crowded RLE or Polygon",
  "url": "/competitions/airbus-ship-detection/discussion/63604",
  "author_name": "",
  "post_date": "2018-08-19T00:05:01.663490Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hey everyone, I have been having trouble finding out what the uncompressed RLE format is and resources talking about it.  I am trying to use detectron, but there is only polygon or crowded RLE format that can be used.</p>\n\n<p>Also is the polygon just the contours of a mask?  Therefore if we have the mask of the encodedPixels then we could just use opencv and get the contours right?</p>",
  "messages": [
    {
      "id": "372320",
      "postDate": "08/19/2018 00:05:01",
      "content": "<p>Hey everyone, I have been having trouble finding out what the uncompressed RLE format is and resources talking about it.  I am trying to use detectron, but there is only polygon or crowded RLE format that can be used.</p>\n\n<p>Also is the polygon just the contours of a mask?  Therefore if we have the mask of the encodedPixels then we could just use opencv and get the contours right?</p>",
      "rawMarkdown": "Hey everyone, I have been having trouble finding out what the uncompressed RLE format is and resources talking about it.  I am trying to use detectron, but there is only polygon or crowded RLE format that can be used.\n\nAlso is the polygon just the contours of a mask?  Therefore if we have the mask of the encodedPixels then we could just use opencv and get the contours right?",
      "votes": null
    },
    {
      "id": "372362",
      "postDate": "08/19/2018 05:07:06",
      "content": "<p>I don't have a 100% grasp the annotations, but I believe you can just take the mask given by the code from other topics in this discussion board and then use opencv to get a polygon for COCO annotations.</p>\n\n<p>I found a post that should help with getting that. \n<a href=\"https://github.com/facebookresearch/Detectron/issues/100\">https://github.com/facebookresearch/Detectron/issues/100</a></p>",
      "rawMarkdown": "I don't have a 100% grasp the annotations, but I believe you can just take the mask given by the code from other topics in this discussion board and then use opencv to get a polygon for COCO annotations.\n\nI found a post that should help with getting that. \nhttps://github.com/facebookresearch/Detectron/issues/100",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 372362,
      "author_name": "bgdev5279",
      "author_url": "",
      "post_date": "08/19/2018 05:07:06",
      "content": "<p>I don't have a 100% grasp the annotations, but I believe you can just take the mask given by the code from other topics in this discussion board and then use opencv to get a polygon for COCO annotations.</p>\n\n<p>I found a post that should help with getting that. \n<a href=\"https://github.com/facebookresearch/Detectron/issues/100\">https://github.com/facebookresearch/Detectron/issues/100</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "372320": "Hey everyone, I have been having trouble finding out what the uncompressed RLE format is and resources talking about it.  I am trying to use detectron, but there is only polygon or crowded RLE format that can be used.\n\nAlso is the polygon just the contours of a mask?  Therefore if we have the mask of the encodedPixels then we could just use opencv and get the contours right?",
    "372362": "I don't have a 100% grasp the annotations, but I believe you can just take the mask given by the code from other topics in this discussion board and then use opencv to get a polygon for COCO annotations.\n\nI found a post that should help with getting that. \nhttps://github.com/facebookresearch/Detectron/issues/100"
  },
  "source": "meta"
}