{
  "id": 54375,
  "title": "new images with only 2 dimensions",
  "url": "/competitions/data-science-bowl-2018/discussion/54375",
  "author_name": "",
  "post_date": "2018-04-12T17:21:09.817631300Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I am kind of confused as to how to make predictions for the new training set. I currently try to take a constant amount of channels for each image, but some of the new images have to channels, and simply have two large dimensions. Does anyone know how we should handle this?</p>",
  "messages": [
    {
      "id": "313028",
      "postDate": "04/12/2018 17:21:09",
      "content": "<p>I am kind of confused as to how to make predictions for the new training set. I currently try to take a constant amount of channels for each image, but some of the new images have to channels, and simply have two large dimensions. Does anyone know how we should handle this?</p>",
      "rawMarkdown": "I am kind of confused as to how to make predictions for the new training set. I currently try to take a constant amount of channels for each image, but some of the new images have to channels, and simply have two large dimensions. Does anyone know how we should handle this?",
      "votes": null
    },
    {
      "id": "313948",
      "postDate": "04/14/2018 05:59:02",
      "content": "<p>Hi, I too had this issue. Just use OpenCV to read the images. This post might be helpful for you\n<a href=\"https://www.kaggle.com/c/data-science-bowl-2018/discussion/54310\">https://www.kaggle.com/c/data-science-bowl-2018/discussion/54310</a></p>",
      "rawMarkdown": "Hi, I too had this issue. Just use OpenCV to read the images. This post might be helpful for you\nhttps://www.kaggle.com/c/data-science-bowl-2018/discussion/54310",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 313948,
      "author_name": "vikramank",
      "author_url": "",
      "post_date": "04/14/2018 05:59:02",
      "content": "<p>Hi, I too had this issue. Just use OpenCV to read the images. This post might be helpful for you\n<a href=\"https://www.kaggle.com/c/data-science-bowl-2018/discussion/54310\">https://www.kaggle.com/c/data-science-bowl-2018/discussion/54310</a></p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "313028": "I am kind of confused as to how to make predictions for the new training set. I currently try to take a constant amount of channels for each image, but some of the new images have to channels, and simply have two large dimensions. Does anyone know how we should handle this?",
    "313948": "Hi, I too had this issue. Just use OpenCV to read the images. This post might be helpful for you\nhttps://www.kaggle.com/c/data-science-bowl-2018/discussion/54310"
  },
  "source": "meta"
}