{
  "id": 32847,
  "title": "Are the winning models generalizable to other locations?",
  "url": "/competitions/dstl-satellite-imagery-feature-detection/discussion/32847",
  "author_name": "Zombie",
  "post_date": "2017-05-11T17:39:35.268000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Just out of interest, would the RGB component to the winning models be generalizable to other locations and data sources. I.e. has anyone tried their model out on some shots from google earth? And if so how did it do?</p>",
  "messages": [
    {
      "id": 181959,
      "postDate": "2017-05-11T17:39:35.270Z",
      "content": "<p>Just out of interest, would the RGB component to the winning models be generalizable to other locations and data sources. I.e. has anyone tried their model out on some shots from google earth? And if so how did it do?</p>",
      "rawMarkdown": "Just out of interest, would the RGB component to the winning models be generalizable to other locations and data sources. I.e. has anyone tried their model out on some shots from google earth? And if so how did it do?"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "181959": "Just out of interest, would the RGB component to the winning models be generalizable to other locations and data sources. I.e. has anyone tried their model out on some shots from google earth? And if so how did it do?"
  }
}