{
  "id": 6507,
  "title": "Negative images to train the classifier",
  "url": "/competitions/dogs-vs-cats/discussion/6507",
  "author_name": "",
  "post_date": "2013-12-03T05:40:59.717Z",
  "votes": null,
  "comment_count": 1,
  "views": 1899,
  "content": "<p>In the data files, I only see dogs and cats images(although there are few images which are neither cats nor dogs but that can be assumed as noise or real human mistakes)</p>\n<p>What about the negative images, which has no dogs and cats and should be used to train the classifier</p>\n<p>Sure, I can use it from some external source but am I missing something or there is another way to approach this problem without using negative images at all?</p>\n<p>Can we train the classifier without negative images?</p>",
  "messages": [
    {
      "id": "35730",
      "postDate": "12/03/2013 05:40:59",
      "content": "<p>In the data files, I only see dogs and cats images(although there are few images which are neither cats nor dogs but that can be assumed as noise or real human mistakes)</p>\n<p>What about the negative images, which has no dogs and cats and should be used to train the classifier</p>\n<p>Sure, I can use it from some external source but am I missing something or there is another way to approach this problem without using negative images at all?</p>\n<p>Can we train the classifier without negative images?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "35824",
      "postDate": "12/04/2013 13:59:00",
      "content": "<p>I don't see why do you need images with no dog or cat. You have 2 classes of objects and the training set of objects each assigned to one of these 2 classes. Obviously the resulting classified will have rather limited applicability as it will be able to identify what object is on the picture, dog or cat, but will work reliably only when the picture with dog or cat is presented.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 35824,
      "author_name": "milakov",
      "author_url": "",
      "post_date": "12/04/2013 13:59:00",
      "content": "<p>I don't see why do you need images with no dog or cat. You have 2 classes of objects and the training set of objects each assigned to one of these 2 classes. Obviously the resulting classified will have rather limited applicability as it will be able to identify what object is on the picture, dog or cat, but will work reliably only when the picture with dog or cat is presented.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "35730": "",
    "35824": ""
  },
  "source": "meta"
}