{
  "id": 16303,
  "title": "Extra images",
  "url": "/competitions/noaa-right-whale-recognition/discussion/16303",
  "author_name": "",
  "post_date": "2015-09-03T12:03:46.517Z",
  "votes": null,
  "comment_count": 3,
  "views": 722,
  "content": "<p>The data page says: &quot;To discourage hand labeling, we have supplemented the test dataset with some images that are resized, cropped, or flipped. These processed images are ignored and don't count towards your score.&quot;</p>\n\n<p>Could anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?</p>\n\n<p>The flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.</p>",
  "messages": [
    {
      "id": "91472",
      "postDate": "09/03/2015 12:03:46",
      "content": "<p>The data page says: &quot;To discourage hand labeling, we have supplemented the test dataset with some images that are resized, cropped, or flipped. These processed images are ignored and don't count towards your score.&quot;</p>\n\n<p>Could anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?</p>\n\n<p>The flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.</p>",
      "rawMarkdown": "The data page says: \"To discourage hand labeling, we have supplemented the test dataset with some images that are resized, cropped, or flipped. These processed images are ignored and don't count towards your score.\"\r\n\r\nCould anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?\r\n\r\nThe flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.",
      "votes": null
    },
    {
      "id": "91478",
      "postDate": "09/03/2015 12:32:30",
      "content": "<p>[quote=Ben Moran;91472]\nCould anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?</p>\n\n<p>The flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.\n[/quote]</p>\n\n<p>Yes, it means that, for example, out of 20 of Test (unlabeled) images, only 15 might actually be scored.</p>\n\n<p>It won't cause problems. Because you can predict any whale you want on those 5 &quot;dummy&quot; images without hurting your score. In other words, it doesn't matter what your algorithm predicts for that image.</p>\n\n<p>It would only (potentially) cause a problem if there were training images that were flipped.</p>",
      "rawMarkdown": "[quote=Ben Moran;91472]\r\nCould anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?\r\n\r\nThe flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.\r\n[/quote]\r\n\r\nYes, it means that, for example, out of 20 of Test (unlabeled) images, only 15 might actually be scored.\r\n\r\nIt won't cause problems. Because you can predict any whale you want on those 5 \"dummy\" images without hurting your score. In other words, it doesn't matter what your algorithm predicts for that image.\r\n\r\nIt would only (potentially) cause a problem if there were training images that were flipped.",
      "votes": null
    },
    {
      "id": "91493",
      "postDate": "09/03/2015 15:13:42",
      "content": "<p>I was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!</p>",
      "rawMarkdown": "I was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!",
      "votes": null
    },
    {
      "id": "91497",
      "postDate": "09/03/2015 15:29:41",
      "content": "<p>[quote=Ben Moran;91493]</p>\n\n<p>I was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!</p>\n\n<p>[/quote]</p>\n\n<p>Ah, got it. Yeah, that makes sense.</p>",
      "rawMarkdown": "[quote=Ben Moran;91493]\r\n\r\nI was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!\r\n\r\n[/quote]\r\n\r\nAh, got it. Yeah, that makes sense.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 91478,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "09/03/2015 12:32:30",
      "content": "<p>[quote=Ben Moran;91472]\nCould anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?</p>\n\n<p>The flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.\n[/quote]</p>\n\n<p>Yes, it means that, for example, out of 20 of Test (unlabeled) images, only 15 might actually be scored.</p>\n\n<p>It won't cause problems. Because you can predict any whale you want on those 5 &quot;dummy&quot; images without hurting your score. In other words, it doesn't matter what your algorithm predicts for that image.</p>\n\n<p>It would only (potentially) cause a problem if there were training images that were flipped.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 91493,
      "author_name": "benmoran",
      "author_url": "",
      "post_date": "09/03/2015 15:13:42",
      "content": "<p>I was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 91497,
      "author_name": "inversion",
      "author_url": "",
      "post_date": "09/03/2015 15:29:41",
      "content": "<p>[quote=Ben Moran;91493]</p>\n\n<p>I was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!</p>\n\n<p>[/quote]</p>\n\n<p>Ah, got it. Yeah, that makes sense.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "91472": "The data page says: \"To discourage hand labeling, we have supplemented the test dataset with some images that are resized, cropped, or flipped. These processed images are ignored and don't count towards your score.\"\r\n\r\nCould anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?\r\n\r\nThe flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.",
    "91478": "[quote=Ben Moran;91472]\r\nCould anyone explain more about what this means?  Are some of the unlabelled pictures in imgs.zip these processed files?\r\n\r\nThe flipping sounds like it could potentially cause problems, if the recognition of individuals depends on asymmetric markings.\r\n[/quote]\r\n\r\nYes, it means that, for example, out of 20 of Test (unlabeled) images, only 15 might actually be scored.\r\n\r\nIt won't cause problems. Because you can predict any whale you want on those 5 \"dummy\" images without hurting your score. In other words, it doesn't matter what your algorithm predicts for that image.\r\n\r\nIt would only (potentially) cause a problem if there were training images that were flipped.",
    "91493": "I was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!",
    "91497": "[quote=Ben Moran;91493]\r\n\r\nI was thinking that it would interfere with semi-supervised approaches that try to make use of the unlabelled data as well.  But thanks!\r\n\r\n[/quote]\r\n\r\nAh, got it. Yeah, that makes sense."
  },
  "source": "meta"
}