{
  "id": 65456,
  "title": "[x,y,w,h] or [y,x,h,w]?",
  "url": "/competitions/rsna-pneumonia-detection-challenge/discussion/65456",
  "author_name": "",
  "post_date": "2018-09-11T07:36:41.160430800Z",
  "votes": 3,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I apologize for the stupid question, but why everywhere the order of coordinates [y,x,h,w] is used instead of [x,y,w,h]?</p>",
  "messages": [
    {
      "id": "385536",
      "postDate": "09/11/2018 07:36:41",
      "content": "<p>I apologize for the stupid question, but why everywhere the order of coordinates [y,x,h,w] is used instead of [x,y,w,h]?</p>",
      "rawMarkdown": "I apologize for the stupid question, but why everywhere the order of coordinates [y,x,h,w] is used instead of [x,y,w,h]?",
      "votes": null
    },
    {
      "id": "385567",
      "postDate": "09/11/2018 08:49:04",
      "content": "<p>I think that's because we use scikit-image library. See <a href=\"http://scikit-image.org/docs/dev/user_guide/numpy_images.html#numpy-images-coordinate-conventions\">http://scikit-image.org/docs/dev/user_guide/numpy_images.html#numpy-images-coordinate-conventions</a></p>\n\n<blockquote>\n  <p>Two-dimensional (2D) grayscale images (such as camera above) are indexed by row and columns (abbreviated to either (row, col) or (r, c))</p>\n</blockquote>",
      "rawMarkdown": "I think that's because we use scikit-image library. See http://scikit-image.org/docs/dev/user_guide/numpy_images.html#numpy-images-coordinate-conventions\n\n&gt; Two-dimensional (2D) grayscale images (such as camera above) are indexed by row and columns (abbreviated to either (row, col) or (r, c))",
      "votes": null
    },
    {
      "id": "385581",
      "postDate": "09/11/2018 09:10:06",
      "content": "<p>It seems to be that the standard way of representing images as 2D numpy array as follows.\nThe first dimension is the height, and the second dimension is the width. The data is ordered by lines, then each line is ordered by pixels.\nFor example, pixel_array of pydicom follows this convention.</p>",
      "rawMarkdown": "It seems to be that the standard way of representing images as 2D numpy array as follows.\nThe first dimension is the height, and the second dimension is the width. The data is ordered by lines, then each line is ordered by pixels.\nFor example, pixel_array of pydicom follows this convention.",
      "votes": null
    },
    {
      "id": "385594",
      "postDate": "09/11/2018 09:27:57",
      "content": "<p>Thank you!</p>",
      "rawMarkdown": "Thank you!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 385567,
      "author_name": "sergeyzlobin",
      "author_url": "",
      "post_date": "09/11/2018 08:49:04",
      "content": "<p>I think that's because we use scikit-image library. See <a href=\"http://scikit-image.org/docs/dev/user_guide/numpy_images.html#numpy-images-coordinate-conventions\">http://scikit-image.org/docs/dev/user_guide/numpy_images.html#numpy-images-coordinate-conventions</a></p>\n\n<blockquote>\n  <p>Two-dimensional (2D) grayscale images (such as camera above) are indexed by row and columns (abbreviated to either (row, col) or (r, c))</p>\n</blockquote>",
      "votes": null,
      "replies": [
        {
          "id": 385581,
          "author_name": "sergeyzlobin",
          "author_url": "",
          "post_date": "09/11/2018 09:10:06",
          "content": "<p>It seems to be that the standard way of representing images as 2D numpy array as follows.\nThe first dimension is the height, and the second dimension is the width. The data is ordered by lines, then each line is ordered by pixels.\nFor example, pixel_array of pydicom follows this convention.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 385594,
          "author_name": "aantonova",
          "author_url": "",
          "post_date": "09/11/2018 09:27:57",
          "content": "<p>Thank you!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "385536": "I apologize for the stupid question, but why everywhere the order of coordinates [y,x,h,w] is used instead of [x,y,w,h]?",
    "385567": "I think that's because we use scikit-image library. See http://scikit-image.org/docs/dev/user_guide/numpy_images.html#numpy-images-coordinate-conventions\n\n&gt; Two-dimensional (2D) grayscale images (such as camera above) are indexed by row and columns (abbreviated to either (row, col) or (r, c))",
    "385581": "It seems to be that the standard way of representing images as 2D numpy array as follows.\nThe first dimension is the height, and the second dimension is the width. The data is ordered by lines, then each line is ordered by pixels.\nFor example, pixel_array of pydicom follows this convention.",
    "385594": "Thank you!"
  },
  "source": "meta"
}