{
  "id": 39271,
  "title": "What's the difference between LOC and detection?",
  "url": "/competitions/imagenet-object-localization-challenge/discussion/39271",
  "author_name": "",
  "post_date": "2017-09-11T01:00:54.473298Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi, I'm new for ImageNet Competition. Sorry for the simple question, but what's the difference between LOC and detection? In the challenge's definition, both of them required to produce bounding boxs and labels, is that right? I see the error equation for LOC, why detection doesn't have it?</p>",
  "messages": [
    {
      "id": "220054",
      "postDate": "09/11/2017 01:00:54",
      "content": "<p>Hi, I'm new for ImageNet Competition. Sorry for the simple question, but what's the difference between LOC and detection? In the challenge's definition, both of them required to produce bounding boxs and labels, is that right? I see the error equation for LOC, why detection doesn't have it?</p>",
      "rawMarkdown": "Hi, I'm new for ImageNet Competition. Sorry for the simple question, but what's the difference between LOC and detection? In the challenge's definition, both of them required to produce bounding boxs and labels, is that right? I see the error equation for LOC, why detection doesn't have it?",
      "votes": null
    },
    {
      "id": "247935",
      "postDate": "11/24/2017 10:57:06",
      "content": "<p>Hi, I do believe the follow (from <a href=\"https://arxiv.org/pdf/1409.0575.pdf\">https://arxiv.org/pdf/1409.0575.pdf</a>):</p>\n\n<p>2.Single-object localization = CLS+LOC\n(2011-2014): Algorithms produce a list of object categories present in the image, along with an axis-aligned bounding box indicating the position and scale of one instance of each object category.</p>\n\n<p>3.Object detection = DET\n(2013-2014): Algorithms produce a list of object categories present in the image along with an axis-aligned bounding box indicating the position and scale of every instance of each object category</p>",
      "rawMarkdown": "Hi, I do believe the follow (from [https://arxiv.org/pdf/1409.0575.pdf][1]):\n\n2.Single-object localization = CLS+LOC\n(2011-2014): Algorithms produce a list of object categories present in the image, along with an axis-aligned bounding box indicating the position and scale of one instance of each object category.\n\n3.Object detection = DET\n(2013-2014): Algorithms produce a list of object categories present in the image along with an axis-aligned bounding box indicating the position and scale of every instance of each object category\n\n\n  [1]: https://arxiv.org/pdf/1409.0575.pdf",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 247935,
      "author_name": "algila",
      "author_url": "",
      "post_date": "11/24/2017 10:57:06",
      "content": "<p>Hi, I do believe the follow (from <a href=\"https://arxiv.org/pdf/1409.0575.pdf\">https://arxiv.org/pdf/1409.0575.pdf</a>):</p>\n\n<p>2.Single-object localization = CLS+LOC\n(2011-2014): Algorithms produce a list of object categories present in the image, along with an axis-aligned bounding box indicating the position and scale of one instance of each object category.</p>\n\n<p>3.Object detection = DET\n(2013-2014): Algorithms produce a list of object categories present in the image along with an axis-aligned bounding box indicating the position and scale of every instance of each object category</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "220054": "Hi, I'm new for ImageNet Competition. Sorry for the simple question, but what's the difference between LOC and detection? In the challenge's definition, both of them required to produce bounding boxs and labels, is that right? I see the error equation for LOC, why detection doesn't have it?",
    "247935": "Hi, I do believe the follow (from [https://arxiv.org/pdf/1409.0575.pdf][1]):\n\n2.Single-object localization = CLS+LOC\n(2011-2014): Algorithms produce a list of object categories present in the image, along with an axis-aligned bounding box indicating the position and scale of one instance of each object category.\n\n3.Object detection = DET\n(2013-2014): Algorithms produce a list of object categories present in the image along with an axis-aligned bounding box indicating the position and scale of every instance of each object category\n\n\n  [1]: https://arxiv.org/pdf/1409.0575.pdf"
  },
  "source": "meta"
}