{
  "id": 121213,
  "title": "Explore image augmentations easily",
  "url": "/competitions/deepfake-detection-challenge/discussion/121213",
  "author_name": "Ilia Larchenko",
  "post_date": "2019-12-11T21:03:31.532000",
  "votes": 12,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi, everybody!\nOne of the essential concepts in modern computer vision that often helps to get better results is image augmentation. And this particular competition should not be an exception - video processing is connected to image processing. I have recently created a tool to explore different image augmentations in a convenient interface. You can check it here: <a href=\"https://albumentations-demo.herokuapp.com/\">https://albumentations-demo.herokuapp.com/</a> or run locally from source <a href=\"https://github.com/IliaLarchenko/albumentations-demo\">https://github.com/IliaLarchenko/albumentations-demo</a></p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F951222%2Fc89bf384e054da45f329ec78a425399a%2Fdemo3.gif?generation=1576097968102467&amp;alt=media\" alt=\"\"></p>\n\n<p>I have also written a short article about augmentations tuning; it can be useful for those who make the first steps in Computer Vision <a href=\"https://towardsdatascience.com/explore-image-augmentations-using-a-convenient-tool-a199b4ac8214\">https://towardsdatascience.com/explore-image-augmentations-using-a-convenient-tool-a199b4ac8214</a></p>\n\n<p>I hope it will help those who participate in this or other CV competitions.</p>",
  "messages": [
    {
      "id": 692915,
      "postDate": "2019-12-11T21:03:31.533Z",
      "content": "<p>Hi, everybody!\nOne of the essential concepts in modern computer vision that often helps to get better results is image augmentation. And this particular competition should not be an exception - video processing is connected to image processing. I have recently created a tool to explore different image augmentations in a convenient interface. You can check it here: <a href=\"https://albumentations-demo.herokuapp.com/\">https://albumentations-demo.herokuapp.com/</a> or run locally from source <a href=\"https://github.com/IliaLarchenko/albumentations-demo\">https://github.com/IliaLarchenko/albumentations-demo</a></p>\n\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F951222%2Fc89bf384e054da45f329ec78a425399a%2Fdemo3.gif?generation=1576097968102467&amp;alt=media\" alt=\"\"></p>\n\n<p>I have also written a short article about augmentations tuning; it can be useful for those who make the first steps in Computer Vision <a href=\"https://towardsdatascience.com/explore-image-augmentations-using-a-convenient-tool-a199b4ac8214\">https://towardsdatascience.com/explore-image-augmentations-using-a-convenient-tool-a199b4ac8214</a></p>\n\n<p>I hope it will help those who participate in this or other CV competitions.</p>",
      "rawMarkdown": "Hi, everybody!\nOne of the essential concepts in modern computer vision that often helps to get better results is image augmentation. And this particular competition should not be an exception - video processing is connected to image processing. I have recently created a tool to explore different image augmentations in a convenient interface. You can check it here: https://albumentations-demo.herokuapp.com/ or run locally from source https://github.com/IliaLarchenko/albumentations-demo\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F951222%2Fc89bf384e054da45f329ec78a425399a%2Fdemo3.gif?generation=1576097968102467&amp;alt=media)\n\n\nI have also written a short article about augmentations tuning; it can be useful for those who make the first steps in Computer Vision https://towardsdatascience.com/explore-image-augmentations-using-a-convenient-tool-a199b4ac8214\n\nI hope it will help those who participate in this or other CV competitions.",
      "votes": 11
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "692915": "Hi, everybody!\nOne of the essential concepts in modern computer vision that often helps to get better results is image augmentation. And this particular competition should not be an exception - video processing is connected to image processing. I have recently created a tool to explore different image augmentations in a convenient interface. You can check it here: https://albumentations-demo.herokuapp.com/ or run locally from source https://github.com/IliaLarchenko/albumentations-demo\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F951222%2Fc89bf384e054da45f329ec78a425399a%2Fdemo3.gif?generation=1576097968102467&amp;alt=media)\n\n\nI have also written a short article about augmentations tuning; it can be useful for those who make the first steps in Computer Vision https://towardsdatascience.com/explore-image-augmentations-using-a-convenient-tool-a199b4ac8214\n\nI hope it will help those who participate in this or other CV competitions."
  }
}