{
  "id": 35741,
  "title": "Dataset",
  "url": "/competitions/passenger-screening-algorithm-challenge/discussion/35741",
  "author_name": "",
  "post_date": "2017-07-04T13:49:33.198389200Z",
  "votes": -3,
  "comment_count": 3,
  "views": 0,
  "content": "<p>May I ask if all of the image data of Stage 1 are \"threat images,\" as opposed to \"non-threat\" images? We need these two classes (threat vs non-threat images) to run AIHPC predictive algos.</p>",
  "messages": [
    {
      "id": "199076",
      "postDate": "07/04/2017 13:49:33",
      "content": "<p>May I ask if all of the image data of Stage 1 are \"threat images,\" as opposed to \"non-threat\" images? We need these two classes (threat vs non-threat images) to run AIHPC predictive algos.</p>",
      "rawMarkdown": "May I ask if all of the image data of Stage 1 are \"threat images,\" as opposed to \"non-threat\" images? We need these two classes (threat vs non-threat images) to run AIHPC predictive algos.",
      "votes": null
    },
    {
      "id": "199117",
      "postDate": "07/04/2017 16:46:47",
      "content": "<p>Each image is divided in 17 zones. On each image some of the zones have threats and other doesn't</p>",
      "rawMarkdown": "Each image is divided in 17 zones. On each image some of the zones have threats and other doesn't",
      "votes": null
    },
    {
      "id": "199135",
      "postDate": "07/04/2017 18:00:47",
      "content": "<p>Thanks. Using zones is not a good idea to solve the problem of detecting threats, as it splits the threats into pieces that may lead to losing sight on the big pictures.</p>",
      "rawMarkdown": "Thanks. Using zones is not a good idea to solve the problem of detecting threats, as it splits the threats into pieces that may lead to losing sight on the big pictures.",
      "votes": null
    },
    {
      "id": "204158",
      "postDate": "07/17/2017 18:33:20",
      "content": "<p>hey DrYe,</p>\n\n<p>The zones are driven by the end use case: the screening personnel only see a cartoon mannequin of the body with a red overlay for a zone with a threat.</p>\n\n<p>Kind of like this: GingerbreadManScanners_full.jpg\"&gt;http://www.popsci.com/sites/popsci.com/files/styles/large_1x_/public/import/2013/images/2011/08/2011_08_12_JA_GingerbreadManScanners_full.jpg</p>\n\n<p>Puneith</p>",
      "rawMarkdown": "hey DrYe,\n\nThe zones are driven by the end use case: the screening personnel only see a cartoon mannequin of the body with a red overlay for a zone with a threat.\n\nKind of like this: http://www.popsci.com/sites/popsci.com/files/styles/large_1x_/public/import/2013/images/2011/08/2011_08_12_JA___GingerbreadManScanners_full.jpg\n\nPuneith",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 199117,
      "author_name": "ironbar",
      "author_url": "",
      "post_date": "07/04/2017 16:46:47",
      "content": "<p>Each image is divided in 17 zones. On each image some of the zones have threats and other doesn't</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 199135,
      "author_name": "drye123",
      "author_url": "",
      "post_date": "07/04/2017 18:00:47",
      "content": "<p>Thanks. Using zones is not a good idea to solve the problem of detecting threats, as it splits the threats into pieces that may lead to losing sight on the big pictures.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 204158,
      "author_name": "puneith",
      "author_url": "",
      "post_date": "07/17/2017 18:33:20",
      "content": "<p>hey DrYe,</p>\n\n<p>The zones are driven by the end use case: the screening personnel only see a cartoon mannequin of the body with a red overlay for a zone with a threat.</p>\n\n<p>Kind of like this: GingerbreadManScanners_full.jpg\"&gt;http://www.popsci.com/sites/popsci.com/files/styles/large_1x_/public/import/2013/images/2011/08/2011_08_12_JA_GingerbreadManScanners_full.jpg</p>\n\n<p>Puneith</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "199076": "May I ask if all of the image data of Stage 1 are \"threat images,\" as opposed to \"non-threat\" images? We need these two classes (threat vs non-threat images) to run AIHPC predictive algos.",
    "199117": "Each image is divided in 17 zones. On each image some of the zones have threats and other doesn't",
    "199135": "Thanks. Using zones is not a good idea to solve the problem of detecting threats, as it splits the threats into pieces that may lead to losing sight on the big pictures.",
    "204158": "hey DrYe,\n\nThe zones are driven by the end use case: the screening personnel only see a cartoon mannequin of the body with a red overlay for a zone with a threat.\n\nKind of like this: http://www.popsci.com/sites/popsci.com/files/styles/large_1x_/public/import/2013/images/2011/08/2011_08_12_JA___GingerbreadManScanners_full.jpg\n\nPuneith"
  },
  "source": "meta"
}