{
  "id": 34443,
  "title": "Cant finish.. so here is where I am...",
  "url": "/competitions/noaa-fisheries-steller-sea-lion-population-count/discussion/34443",
  "author_name": "karle",
  "post_date": "2017-06-09T07:49:58.172000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Here is what I came so with so far. I can't finish (run times), a wedding, and business trip next week...\nI took a pragmatic approach to identify features.\nI have two programs, the one included, and the one that is not ready for prime time.\nThe program takes two matched files ( train and dotted) and creates 5 files.</p>\n\n<ul>\n<li>gray scale Train</li>\n<li>gray scale Blur (which becomes the hints for all seal lions)</li>\n<li>gray scale Edges</li>\n<li>gray scale Dots (just the dots , all else white)</li>\n<li>One ASCII file of the images, format:  T, E, B, D</li>\n</ul>\n\n<p>What I think are novel features are :\n - I swap the red and green channels on Train to make green sea lions (false color)\n - use fast gray scale by just using green channel (no float math) - sea lions pop!\n - create 4 new channels (T, E, B, D)\n - use a sliding windows on the stacked 4 channels  to count sea lions  (not included)\n - create a database of characteristics of sea lions found (not included)</p>\n\n<p>Take a look at the files (all based on 0.jpg) \nLet me know what you think about this  feature exploration ... not a solution!</p>",
  "messages": [
    {
      "id": 191064,
      "postDate": "2017-06-09T07:49:58.173Z",
      "content": "<p>Here is what I came so with so far. I can't finish (run times), a wedding, and business trip next week...\nI took a pragmatic approach to identify features.\nI have two programs, the one included, and the one that is not ready for prime time.\nThe program takes two matched files ( train and dotted) and creates 5 files.</p>\n\n<ul>\n<li>gray scale Train</li>\n<li>gray scale Blur (which becomes the hints for all seal lions)</li>\n<li>gray scale Edges</li>\n<li>gray scale Dots (just the dots , all else white)</li>\n<li>One ASCII file of the images, format:  T, E, B, D</li>\n</ul>\n\n<p>What I think are novel features are :\n - I swap the red and green channels on Train to make green sea lions (false color)\n - use fast gray scale by just using green channel (no float math) - sea lions pop!\n - create 4 new channels (T, E, B, D)\n - use a sliding windows on the stacked 4 channels  to count sea lions  (not included)\n - create a database of characteristics of sea lions found (not included)</p>\n\n<p>Take a look at the files (all based on 0.jpg) \nLet me know what you think about this  feature exploration ... not a solution!</p>",
      "rawMarkdown": "Here is what I came so with so far. I can't finish (run times), a wedding, and business trip next week...\nI took a pragmatic approach to identify features.\nI have two programs, the one included, and the one that is not ready for prime time.\nThe program takes two matched files ( train and dotted) and creates 5 files.\n\n - gray scale Train\n - gray scale Blur (which becomes the hints for all seal lions)\n - gray scale Edges\n - gray scale Dots (just the dots , all else white)\n - One ASCII file of the images, format:  T, E, B, D\n\nWhat I think are novel features are :\n - I swap the red and green channels on Train to make green sea lions (false color)\n - use fast gray scale by just using green channel (no float math) - sea lions pop!\n - create 4 new channels (T, E, B, D)\n - use a sliding windows on the stacked 4 channels  to count sea lions  (not included)\n - create a database of characteristics of sea lions found (not included)\n\nTake a look at the files (all based on 0.jpg) \nLet me know what you think about this  feature exploration ... not a solution!",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "191064": "Here is what I came so with so far. I can't finish (run times), a wedding, and business trip next week...\nI took a pragmatic approach to identify features.\nI have two programs, the one included, and the one that is not ready for prime time.\nThe program takes two matched files ( train and dotted) and creates 5 files.\n\n - gray scale Train\n - gray scale Blur (which becomes the hints for all seal lions)\n - gray scale Edges\n - gray scale Dots (just the dots , all else white)\n - One ASCII file of the images, format:  T, E, B, D\n\nWhat I think are novel features are :\n - I swap the red and green channels on Train to make green sea lions (false color)\n - use fast gray scale by just using green channel (no float math) - sea lions pop!\n - create 4 new channels (T, E, B, D)\n - use a sliding windows on the stacked 4 channels  to count sea lions  (not included)\n - create a database of characteristics of sea lions found (not included)\n\nTake a look at the files (all based on 0.jpg) \nLet me know what you think about this  feature exploration ... not a solution!"
  }
}