{
  "id": 177873,
  "title": "Optical flow in motion prediction",
  "url": "/competitions/lyft-motion-prediction-autonomous-vehicles/discussion/177873",
  "author_name": "",
  "post_date": "2020-08-27T16:44:05.941651300Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Motion prediction often involves “optical flow” algorithms. The flow vectors are used to predict direction and speed based on lateral motion.</p>\n<p>OpenCV has many state-of-the-art optical flow implementations - both for dense and sparse features.(<a href=\"https://docs.opencv.org/master/d4/dee/tutorial_optical_flow.html\" target=\"_blank\">https://docs.opencv.org/master/d4/dee/tutorial_optical_flow.html</a>)</p>\n<p>Many CNN networks like Flownet also accurately estimate optical flow accurately, even in the presence of occlusions.</p>\n<p>For more optical flow techniques, you can go through this page (<a href=\"https://github.com/LearnCV/awesome-optical-flow-algorithm\" target=\"_blank\">https://github.com/LearnCV/awesome-optical-flow-algorithm</a> ). It has links to paper and code to many such algorithms.</p>",
  "messages": [
    {
      "id": "987955",
      "postDate": "08/27/2020 16:44:05",
      "content": "<p>Motion prediction often involves “optical flow” algorithms. The flow vectors are used to predict direction and speed based on lateral motion.</p>\n<p>OpenCV has many state-of-the-art optical flow implementations - both for dense and sparse features.(<a href=\"https://docs.opencv.org/master/d4/dee/tutorial_optical_flow.html\" target=\"_blank\">https://docs.opencv.org/master/d4/dee/tutorial_optical_flow.html</a>)</p>\n<p>Many CNN networks like Flownet also accurately estimate optical flow accurately, even in the presence of occlusions.</p>\n<p>For more optical flow techniques, you can go through this page (<a href=\"https://github.com/LearnCV/awesome-optical-flow-algorithm\" target=\"_blank\">https://github.com/LearnCV/awesome-optical-flow-algorithm</a> ). It has links to paper and code to many such algorithms.</p>",
      "rawMarkdown": "Motion prediction often involves “optical flow” algorithms. The flow vectors are used to predict direction and speed based on lateral motion.\n\nOpenCV has many state-of-the-art optical flow implementations - both for dense and sparse features.(https://docs.opencv.org/master/d4/dee/tutorial_optical_flow.html)\n\nMany CNN networks like Flownet also accurately estimate optical flow accurately, even in the presence of occlusions.\n\nFor more optical flow techniques, you can go through this page (https://github.com/LearnCV/awesome-optical-flow-algorithm ). It has links to paper and code to many such algorithms.",
      "votes": null
    },
    {
      "id": "989618",
      "postDate": "08/29/2020 00:46:35",
      "content": "<p>I'm not an expert in DL but this is very helpful. Thanks.</p>",
      "rawMarkdown": "I'm not an expert in DL but this is very helpful. Thanks.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 989618,
      "author_name": "rishidamarla",
      "author_url": "",
      "post_date": "08/29/2020 00:46:35",
      "content": "<p>I'm not an expert in DL but this is very helpful. Thanks.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "987955": "Motion prediction often involves “optical flow” algorithms. The flow vectors are used to predict direction and speed based on lateral motion.\n\nOpenCV has many state-of-the-art optical flow implementations - both for dense and sparse features.(https://docs.opencv.org/master/d4/dee/tutorial_optical_flow.html)\n\nMany CNN networks like Flownet also accurately estimate optical flow accurately, even in the presence of occlusions.\n\nFor more optical flow techniques, you can go through this page (https://github.com/LearnCV/awesome-optical-flow-algorithm ). It has links to paper and code to many such algorithms.",
    "989618": "I'm not an expert in DL but this is very helpful. Thanks."
  },
  "source": "meta"
}