{
  "id": 51469,
  "title": "Caffe dice loss layer for 2D images",
  "url": "/competitions/data-science-bowl-2018/discussion/51469",
  "author_name": "Rishabh",
  "post_date": "2018-03-09T08:31:11.577000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi Kaggle Community,</p>\n\n<p>I have made a dice loss layer in Caffe that works well with highly imbalanced images in which there is very less foreground in comparison to the background.</p>\n\n<p>After seeing this competition's data, I feel it will be useful to people who are using Caffe for this competition. </p>\n\n<p>Please find the implementation at the link. <a href=\"https://github.com/im-rishabh/Caffe-Dice-Loss-Layer/blob/master/dice_loss.py\">Dice Loss Implementation in Caffe</a></p>",
  "messages": [
    {
      "id": 293136,
      "postDate": "2018-03-09T08:31:11.577Z",
      "content": "<p>Hi Kaggle Community,</p>\n\n<p>I have made a dice loss layer in Caffe that works well with highly imbalanced images in which there is very less foreground in comparison to the background.</p>\n\n<p>After seeing this competition's data, I feel it will be useful to people who are using Caffe for this competition. </p>\n\n<p>Please find the implementation at the link. <a href=\"https://github.com/im-rishabh/Caffe-Dice-Loss-Layer/blob/master/dice_loss.py\">Dice Loss Implementation in Caffe</a></p>",
      "rawMarkdown": "Hi Kaggle Community,\n\nI have made a dice loss layer in Caffe that works well with highly imbalanced images in which there is very less foreground in comparison to the background.\n\nAfter seeing this competition's data, I feel it will be useful to people who are using Caffe for this competition. \n\nPlease find the implementation at the link. [Dice Loss Implementation in Caffe][1]\n\n\n  [1]: https://github.com/im-rishabh/Caffe-Dice-Loss-Layer/blob/master/dice_loss.py"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "293136": "Hi Kaggle Community,\n\nI have made a dice loss layer in Caffe that works well with highly imbalanced images in which there is very less foreground in comparison to the background.\n\nAfter seeing this competition's data, I feel it will be useful to people who are using Caffe for this competition. \n\nPlease find the implementation at the link. [Dice Loss Implementation in Caffe][1]\n\n\n  [1]: https://github.com/im-rishabh/Caffe-Dice-Loss-Layer/blob/master/dice_loss.py"
  }
}