{
  "id": 20166,
  "title": "Neural networks vs CV !",
  "url": "/competitions/state-farm-distracted-driver-detection/discussion/20166",
  "author_name": "",
  "post_date": "2016-04-16T06:28:48.113Z",
  "votes": null,
  "comment_count": 3,
  "views": 1006,
  "content": "<p>Hi, \nThe question might seem a little preliminary. But I was wondering why people are not using Neural networks? Are there any specific advantages of using OpenCV over neural networks? Has somebody tried it out?</p>",
  "messages": [
    {
      "id": "115089",
      "postDate": "04/16/2016 06:28:48",
      "content": "<p>Hi, \nThe question might seem a little preliminary. But I was wondering why people are not using Neural networks? Are there any specific advantages of using OpenCV over neural networks? Has somebody tried it out?</p>",
      "rawMarkdown": "Hi, \r\nThe question might seem a little preliminary. But I was wondering why people are not using Neural networks? Are there any specific advantages of using OpenCV over neural networks? Has somebody tried it out?",
      "votes": null
    },
    {
      "id": "115095",
      "postDate": "04/16/2016 07:27:59",
      "content": "<p>Where did you get that impression?</p>\n\n<p>Pretty sure everyone is using neural networks for this competition.</p>",
      "rawMarkdown": "Where did you get that impression?\r\n\r\nPretty sure everyone is using neural networks for this competition.",
      "votes": null
    },
    {
      "id": "115098",
      "postDate": "04/16/2016 07:46:57",
      "content": "<p>The discussion about OpenCV is around reading image files and converting to data input for NNs. It is also a useful tool for augmenting the data set (e.g. rotations, skews, contrast changes) and/or normalising it. Now we are allowed external data, that could also include detecting key objects in each image via the Haar cascades.</p>",
      "rawMarkdown": "The discussion about OpenCV is around reading image files and converting to data input for NNs. It is also a useful tool for augmenting the data set (e.g. rotations, skews, contrast changes) and/or normalising it. Now we are allowed external data, that could also include detecting key objects in each image via the Haar cascades.",
      "votes": null
    },
    {
      "id": "115100",
      "postDate": "04/16/2016 07:50:34",
      "content": "<p>Thanks Neil and Roman, \nThat clarifies the doubts. I asked the question in haste without going through all the forum posts. I apparently read too much into the openCV part and completely missed the NN  threads.</p>",
      "rawMarkdown": "Thanks Neil and Roman, \r\nThat clarifies the doubts. I asked the question in haste without going through all the forum posts. I apparently read too much into the openCV part and completely missed the NN  threads.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 115095,
      "author_name": "inoryy",
      "author_url": "",
      "post_date": "04/16/2016 07:27:59",
      "content": "<p>Where did you get that impression?</p>\n\n<p>Pretty sure everyone is using neural networks for this competition.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 115098,
      "author_name": "slobo777",
      "author_url": "",
      "post_date": "04/16/2016 07:46:57",
      "content": "<p>The discussion about OpenCV is around reading image files and converting to data input for NNs. It is also a useful tool for augmenting the data set (e.g. rotations, skews, contrast changes) and/or normalising it. Now we are allowed external data, that could also include detecting key objects in each image via the Haar cascades.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 115100,
      "author_name": "mkkori",
      "author_url": "",
      "post_date": "04/16/2016 07:50:34",
      "content": "<p>Thanks Neil and Roman, \nThat clarifies the doubts. I asked the question in haste without going through all the forum posts. I apparently read too much into the openCV part and completely missed the NN  threads.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "115089": "Hi, \r\nThe question might seem a little preliminary. But I was wondering why people are not using Neural networks? Are there any specific advantages of using OpenCV over neural networks? Has somebody tried it out?",
    "115095": "Where did you get that impression?\r\n\r\nPretty sure everyone is using neural networks for this competition.",
    "115098": "The discussion about OpenCV is around reading image files and converting to data input for NNs. It is also a useful tool for augmenting the data set (e.g. rotations, skews, contrast changes) and/or normalising it. Now we are allowed external data, that could also include detecting key objects in each image via the Haar cascades.",
    "115100": "Thanks Neil and Roman, \r\nThat clarifies the doubts. I asked the question in haste without going through all the forum posts. I apparently read too much into the openCV part and completely missed the NN  threads."
  },
  "source": "meta"
}