{
  "id": 121258,
  "title": "Whats the best way to handle data?",
  "url": "/competitions/deepfake-detection-challenge/discussion/121258",
  "author_name": "",
  "post_date": "2019-12-12T04:05:02.130666Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>What is the best way to handle the data? I've written a script to convert the videos into numpy arrays, but it would take hours to convert even a fraction of the data and the files become even larger so saving isnt an option.</p>",
  "messages": [
    {
      "id": "693127",
      "postDate": "12/12/2019 04:05:02",
      "content": "<p>What is the best way to handle the data? I've written a script to convert the videos into numpy arrays, but it would take hours to convert even a fraction of the data and the files become even larger so saving isnt an option.</p>",
      "rawMarkdown": "What is the best way to handle the data? I've written a script to convert the videos into numpy arrays, but it would take hours to convert even a fraction of the data and the files become even larger so saving isnt an option.",
      "votes": null
    },
    {
      "id": "693518",
      "postDate": "12/12/2019 13:19:14",
      "content": "<p>Maybe we just need the region of interest(face) rather than full fraction <a href=\"/gregoryeales\">@gregoryeales</a> </p>",
      "rawMarkdown": "Maybe we just need the region of interest(face) rather than full fraction @gregoryeales",
      "votes": null
    },
    {
      "id": "693520",
      "postDate": "12/12/2019 13:19:59",
      "content": "<p>look here.\n<a href=\"https://www.kaggle.com/marcovasquez/basic-eda-face-detection-split-video-roi\">https://www.kaggle.com/marcovasquez/basic-eda-face-detection-split-video-roi</a></p>",
      "rawMarkdown": "look here.\nhttps://www.kaggle.com/marcovasquez/basic-eda-face-detection-split-video-roi",
      "votes": null
    },
    {
      "id": "693762",
      "postDate": "12/12/2019 19:09:47",
      "content": "<p>Yes, I actually have the same method of loading the image into an array using cv2. I don't know if I like the idea of segmenting the image much.</p>",
      "rawMarkdown": "Yes, I actually have the same method of loading the image into an array using cv2. I don't know if I like the idea of segmenting the image much.",
      "votes": null
    },
    {
      "id": "693765",
      "postDate": "12/12/2019 19:15:42",
      "content": "<p>What would you suggest doing if there are multiple people in the image?</p>",
      "rawMarkdown": "What would you suggest doing if there are multiple people in the image?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 693518,
      "author_name": "diegojohnson",
      "author_url": "",
      "post_date": "12/12/2019 13:19:14",
      "content": "<p>Maybe we just need the region of interest(face) rather than full fraction <a href=\"/gregoryeales\">@gregoryeales</a> </p>",
      "votes": null,
      "replies": [
        {
          "id": 693765,
          "author_name": "gregoryeales",
          "author_url": "",
          "post_date": "12/12/2019 19:15:42",
          "content": "<p>What would you suggest doing if there are multiple people in the image?</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 693520,
      "author_name": "diegojohnson",
      "author_url": "",
      "post_date": "12/12/2019 13:19:59",
      "content": "<p>look here.\n<a href=\"https://www.kaggle.com/marcovasquez/basic-eda-face-detection-split-video-roi\">https://www.kaggle.com/marcovasquez/basic-eda-face-detection-split-video-roi</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 693762,
          "author_name": "gregoryeales",
          "author_url": "",
          "post_date": "12/12/2019 19:09:47",
          "content": "<p>Yes, I actually have the same method of loading the image into an array using cv2. I don't know if I like the idea of segmenting the image much.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "693127": "What is the best way to handle the data? I've written a script to convert the videos into numpy arrays, but it would take hours to convert even a fraction of the data and the files become even larger so saving isnt an option.",
    "693518": "Maybe we just need the region of interest(face) rather than full fraction @gregoryeales",
    "693520": "look here.\nhttps://www.kaggle.com/marcovasquez/basic-eda-face-detection-split-video-roi",
    "693762": "Yes, I actually have the same method of loading the image into an array using cv2. I don't know if I like the idea of segmenting the image much.",
    "693765": "What would you suggest doing if there are multiple people in the image?"
  },
  "source": "meta"
}