{
  "id": 144523,
  "title": "Deepfake Detection models for detect COVID-19 (only question) ?",
  "url": "/competitions/deepfake-detection-challenge/discussion/144523",
  "author_name": "Anatoly Volodko",
  "post_date": "2020-04-19T12:38:22.769000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Probably a small chance, but may be it makes sense to try to apply Deepfake Detection Challenge  models to identify people infected with coronavirus?</p>\n\n<p>For example \n- input: 10000 photos of virus-free people (faces), 10000 photos of infected people\n- labels: 1 - infected, 0 - uninfected</p>\n\n<p>May be instead photos of faces - photos of eyes.</p>\n\n<p>There are no good words about iridology in <a href=\"https://en.wikipedia.org/wiki/Iridology\">wikipedia</a> but may be it's worth trying to apply convolution neural networks in this direction ?</p>",
  "messages": [
    {
      "id": 813163,
      "postDate": "2020-04-19T12:38:22.770Z",
      "content": "<p>Probably a small chance, but may be it makes sense to try to apply Deepfake Detection Challenge  models to identify people infected with coronavirus?</p>\n\n<p>For example \n- input: 10000 photos of virus-free people (faces), 10000 photos of infected people\n- labels: 1 - infected, 0 - uninfected</p>\n\n<p>May be instead photos of faces - photos of eyes.</p>\n\n<p>There are no good words about iridology in <a href=\"https://en.wikipedia.org/wiki/Iridology\">wikipedia</a> but may be it's worth trying to apply convolution neural networks in this direction ?</p>",
      "rawMarkdown": "Probably a small chance, but may be it makes sense to try to apply Deepfake Detection Challenge  models to identify people infected with coronavirus?\n\nFor example \n- input: 10000 photos of virus-free people (faces), 10000 photos of infected people\n- labels: 1 - infected, 0 - uninfected\n\nMay be instead photos of faces - photos of eyes.\n\nThere are no good words about iridology in [wikipedia](https://en.wikipedia.org/wiki/Iridology) but may be it's worth trying to apply convolution neural networks in this direction ?"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "813163": "Probably a small chance, but may be it makes sense to try to apply Deepfake Detection Challenge  models to identify people infected with coronavirus?\n\nFor example \n- input: 10000 photos of virus-free people (faces), 10000 photos of infected people\n- labels: 1 - infected, 0 - uninfected\n\nMay be instead photos of faces - photos of eyes.\n\nThere are no good words about iridology in [wikipedia](https://en.wikipedia.org/wiki/Iridology) but may be it's worth trying to apply convolution neural networks in this direction ?"
  }
}