{
  "id": 12675,
  "title": "Variety of Image Conditions",
  "url": "/competitions/diabetic-retinopathy-detection/discussion/12675",
  "author_name": "",
  "post_date": "2015-03-03T23:08:05.553Z",
  "votes": 2,
  "comment_count": 1,
  "views": 1382,
  "content": "<p>Looking at the size of the photos, there seems to be 4 normal distributions potentially indicating 4 unique conditions by which the photos were taken. Curious how&nbsp;others are sampling. I've&nbsp;considered using&nbsp;reservoir sampling, but would like to hear what you think.</p>",
  "messages": [
    {
      "id": "65370",
      "postDate": "03/03/2015 23:08:05",
      "content": "<p>Looking at the size of the photos, there seems to be 4 normal distributions potentially indicating 4 unique conditions by which the photos were taken. Curious how&nbsp;others are sampling. I've&nbsp;considered using&nbsp;reservoir sampling, but would like to hear what you think.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "66048",
      "postDate": "03/12/2015 15:57:16",
      "content": "<p>Gideon: I find 29 different image sizes if I look at pixel counts (width * height). &nbsp;Plotting the file size distributions grouped by pixel counts makes them all look normal again, I think. &nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 66048,
      "author_name": "paulperry",
      "author_url": "",
      "post_date": "03/12/2015 15:57:16",
      "content": "<p>Gideon: I find 29 different image sizes if I look at pixel counts (width * height). &nbsp;Plotting the file size distributions grouped by pixel counts makes them all look normal again, I think. &nbsp;</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "65370": "",
    "66048": ""
  },
  "source": "meta"
}