{
  "id": 62382,
  "title": "Object Detection and Classification",
  "url": "/competitions/google-ai-open-images-object-detection-track/discussion/62382",
  "author_name": "",
  "post_date": "2018-08-01T02:18:00.047689200Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi  Quick quaestion. Does the neural network beside detection the objects  abd boxing them there is a need to classify them according to 600 class provided by the dataset? I think I am missing something here. Please provide ur input? Thx</p>",
  "messages": [
    {
      "id": "364664",
      "postDate": "08/01/2018 02:18:00",
      "content": "<p>Hi  Quick quaestion. Does the neural network beside detection the objects  abd boxing them there is a need to classify them according to 600 class provided by the dataset? I think I am missing something here. Please provide ur input? Thx</p>",
      "rawMarkdown": "Hi  Quick quaestion. Does the neural network beside detection the objects  abd boxing them there is a need to classify them according to 600 class provided by the dataset? I think I am missing something here. Please provide ur input? Thx",
      "votes": null
    },
    {
      "id": "364667",
      "postDate": "08/01/2018 02:29:33",
      "content": "<p>Short answer... Yes. The submission is all bounding boxes and classes of objects in the test set</p>",
      "rawMarkdown": "Short answer... Yes. The submission is all bounding boxes and classes of objects in the test set",
      "votes": null
    },
    {
      "id": "369078",
      "postDate": "08/11/2018 22:00:39",
      "content": "<p>Hi.  I started to work with data and I realized that I am only focusing on data in train-annotations-bbox and class-descriptions-boxable. I am not very sure about the data in the files: train-annotations-human-imagelabels, and train-annotations-human-imagelabels-boxable and train-annotations-human-imagelabels-boxable. Would u please provide some clarification about whether my approach is not complete. Please also confirm teh deadline for augest 23 rd or whether there might be an extension. Thx u for ur support Regards</p>",
      "rawMarkdown": "Hi.  I started to work with data and I realized that I am only focusing on data in train-annotations-bbox and class-descriptions-boxable. I am not very sure about the data in the files: train-annotations-human-imagelabels, and train-annotations-human-imagelabels-boxable and train-annotations-human-imagelabels-boxable. Would u please provide some clarification about whether my approach is not complete. Please also confirm teh deadline for augest 23 rd or whether there might be an extension. Thx u for ur support Regards",
      "votes": null
    },
    {
      "id": "369332",
      "postDate": "08/12/2018 17:56:57",
      "content": "<ul>\n<li><p>train-annotations-human-imagelabels contains the image-level labels (no boxes) for the full set of images where we have image-level annotations (~5M).</p></li>\n<li><p>train-annotations-human-imagelabels-boxable contains the image-level labels (no boxes) for the subset of images where we have boxes (~1.7M). </p></li>\n</ul>\n\n<p>Whether to use or not the image-level labels is up to you.</p>\n\n<p>Deadline is final, no extensions will be given, sorry.</p>",
      "rawMarkdown": "train-annotations-human-imagelabels contains the image-level labels (no boxes) for the full set of images where we have image-level annotations (~5M).\n\n- train-annotations-human-imagelabels-boxable contains the image-level labels (no boxes) for the subset of images where we have boxes (~1.7M). \n\nWhether to use or not the image-level labels is up to you.\n\nDeadline is final, no extensions will be given, sorry.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 364667,
      "author_name": "moshel",
      "author_url": "",
      "post_date": "08/01/2018 02:29:33",
      "content": "<p>Short answer... Yes. The submission is all bounding boxes and classes of objects in the test set</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 369078,
      "author_name": "hadielzein",
      "author_url": "",
      "post_date": "08/11/2018 22:00:39",
      "content": "<p>Hi.  I started to work with data and I realized that I am only focusing on data in train-annotations-bbox and class-descriptions-boxable. I am not very sure about the data in the files: train-annotations-human-imagelabels, and train-annotations-human-imagelabels-boxable and train-annotations-human-imagelabels-boxable. Would u please provide some clarification about whether my approach is not complete. Please also confirm teh deadline for augest 23 rd or whether there might be an extension. Thx u for ur support Regards</p>",
      "votes": null,
      "replies": [
        {
          "id": 369332,
          "author_name": "jponttusset",
          "author_url": "",
          "post_date": "08/12/2018 17:56:57",
          "content": "<ul>\n<li><p>train-annotations-human-imagelabels contains the image-level labels (no boxes) for the full set of images where we have image-level annotations (~5M).</p></li>\n<li><p>train-annotations-human-imagelabels-boxable contains the image-level labels (no boxes) for the subset of images where we have boxes (~1.7M). </p></li>\n</ul>\n\n<p>Whether to use or not the image-level labels is up to you.</p>\n\n<p>Deadline is final, no extensions will be given, sorry.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "364664": "Hi  Quick quaestion. Does the neural network beside detection the objects  abd boxing them there is a need to classify them according to 600 class provided by the dataset? I think I am missing something here. Please provide ur input? Thx",
    "364667": "Short answer... Yes. The submission is all bounding boxes and classes of objects in the test set",
    "369078": "Hi.  I started to work with data and I realized that I am only focusing on data in train-annotations-bbox and class-descriptions-boxable. I am not very sure about the data in the files: train-annotations-human-imagelabels, and train-annotations-human-imagelabels-boxable and train-annotations-human-imagelabels-boxable. Would u please provide some clarification about whether my approach is not complete. Please also confirm teh deadline for augest 23 rd or whether there might be an extension. Thx u for ur support Regards",
    "369332": "train-annotations-human-imagelabels contains the image-level labels (no boxes) for the full set of images where we have image-level annotations (~5M).\n\n- train-annotations-human-imagelabels-boxable contains the image-level labels (no boxes) for the subset of images where we have boxes (~1.7M). \n\nWhether to use or not the image-level labels is up to you.\n\nDeadline is final, no extensions will be given, sorry."
  },
  "source": "meta"
}