{
  "id": 93252,
  "title": "Should we detect every apparel object in test images?",
  "url": "/competitions/imaterialist-fashion-2019-FGVC6/discussion/93252",
  "author_name": "",
  "post_date": "2019-05-24T16:42:46.154494500Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>According to the labeling on train data, looks like we only have to detect the apparel objects associated with the main person. Just want to make sure about that. \nI borrow the image from Pednoi's kernel <a href=\"https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07\">https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07</a>, which is a great kernel!  Thanks for sharing. <a href=\"/pednoi\">@pednoi</a> </p>\n\n<p>From this image, are the man's pants on the right within the scope of the evaluation?</p>",
  "messages": [
    {
      "id": "536549",
      "postDate": "05/24/2019 16:42:46",
      "content": "<p>According to the labeling on train data, looks like we only have to detect the apparel objects associated with the main person. Just want to make sure about that. \nI borrow the image from Pednoi's kernel <a href=\"https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07\">https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07</a>, which is a great kernel!  Thanks for sharing. <a href=\"/pednoi\">@pednoi</a> </p>\n\n<p>From this image, are the man's pants on the right within the scope of the evaluation?</p>",
      "rawMarkdown": "According to the labeling on train data, looks like we only have to detect the apparel objects associated with the main person. Just want to make sure about that. \nI borrow the image from Pednoi's kernel https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07, which is a great kernel!  Thanks for sharing. @pednoi \n\nFrom this image, are the man's pants on the right within the scope of the evaluation?",
      "votes": null
    },
    {
      "id": "536621",
      "postDate": "05/24/2019 20:38:18",
      "content": "<p>Hi <a href=\"/cooleel\">@cooleel</a>, you are right. for images with multiple people, we only consider the primary person in the image. </p>",
      "rawMarkdown": "Hi @cooleel, you are right. for images with multiple people, we only consider the primary person in the image.",
      "votes": null
    },
    {
      "id": "536658",
      "postDate": "05/24/2019 23:34:03",
      "content": "<p>Great, thank you for the clarification.</p>",
      "rawMarkdown": "Great, thank you for the clarification.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 536621,
      "author_name": "makeitworkjml",
      "author_url": "",
      "post_date": "05/24/2019 20:38:18",
      "content": "<p>Hi <a href=\"/cooleel\">@cooleel</a>, you are right. for images with multiple people, we only consider the primary person in the image. </p>",
      "votes": null,
      "replies": [
        {
          "id": 536658,
          "author_name": "cooleel",
          "author_url": "",
          "post_date": "05/24/2019 23:34:03",
          "content": "<p>Great, thank you for the clarification.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "536549": "According to the labeling on train data, looks like we only have to detect the apparel objects associated with the main person. Just want to make sure about that. \nI borrow the image from Pednoi's kernel https://www.kaggle.com/pednoi/training-mask-r-cnn-to-be-a-fashionista-lb-0-07, which is a great kernel!  Thanks for sharing. @pednoi \n\nFrom this image, are the man's pants on the right within the scope of the evaluation?",
    "536621": "Hi @cooleel, you are right. for images with multiple people, we only consider the primary person in the image.",
    "536658": "Great, thank you for the clarification."
  },
  "source": "meta"
}