{
  "id": 147688,
  "title": "Mask with no value - issue with Mask",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/147688",
  "author_name": "yuvaramsingh",
  "post_date": "2020-05-01T12:23:26.340000",
  "votes": 0,
  "comment_count": 6,
  "views": 0,
  "content": "<p>During my EDA i came across few mask images which has no nonzero values, which means it has no useful information that could be of use . There are only 4 such mask which is not having any values </p>\n\n<p>out of this, <code>3790f55cad63053e956fb73027179707</code> is just a white image with no sample on it. </p>\n\n<p>&gt; 4a2ca53f240932e46eaf8959cb3f490a\n3790f55cad63053e956fb73027179707\ne4215cfc8c41ec04a55431cc413688a9\naaa5732cd49bffddf0d2b7d36fbb0a83</p>\n\n<p>Code to check\n```\nimport numpy as np\nimport skimage.io</p>\n\n<p>image_id = \"aaa5732cd49bffddf0d2b7d36fbb0a83\"</p>\n\n<p>img_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_images/\" + image_id + \".tiff\"\nmask_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_label_masks/\" + image_id \n+ \"_mask.tiff\"\nim = skimage.io.MultiImage(img_path)\nim_mask = skimage.io.MultiImage(mask_path)</p>\n\n<p>non_zero = np.nonzero(im_mask[-1])\nprint(\"img non zero values : \",non_zero)\n```</p>",
  "messages": [
    {
      "id": 830982,
      "postDate": "2020-05-03T02:47:49.597Z",
      "content": "<p>Added this note to <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/148060\">my thread tracking all odd slides</a>. If you come across more let me know and I can keep the running list updated!</p>",
      "rawMarkdown": "Added this note to [my thread tracking all odd slides](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/148060). If you come across more let me know and I can keep the running list updated!",
      "votes": 1,
      "replies": [
        {
          "id": 831020,
          "postDate": "2020-05-03T03:36:52.310Z",
          "content": "<p><a href=\"https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data\">This is a list of more than 100 images waiting to be deleted</a></p>",
          "rawMarkdown": "[This is a list of more than 100 images waiting to be deleted](https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data)",
          "votes": 1
        },
        {
          "id": 831537,
          "postDate": "2020-05-03T12:25:55.650Z",
          "content": "<p>Great thanks I added that to the discussion! </p>",
          "rawMarkdown": "Great thanks I added that to the discussion! "
        },
        {
          "id": 832278,
          "postDate": "2020-05-04T02:40:05.950Z",
          "content": "<p>sure  <a href=\"/dannellyz\">@dannellyz</a> </p>",
          "rawMarkdown": "sure  @dannellyz "
        }
      ]
    },
    {
      "id": 829993,
      "postDate": "2020-05-02T08:31:23.780Z",
      "content": "<p>Oh yeah, there is also a biopsy that is fully white. That the way it is with large datasets ;) </p>",
      "rawMarkdown": "Oh yeah, there is also a biopsy that is fully white. That the way it is with large datasets ;) ",
      "votes": 1
    },
    {
      "id": 829011,
      "postDate": "2020-05-01T12:47:43.993Z",
      "content": "<p>There is such  mask, which can be deleted or used as a verification set during training</p>",
      "rawMarkdown": "There is such  mask, which can be deleted or used as a verification set during training",
      "votes": 1
    },
    {
      "id": 828983,
      "postDate": "2020-05-01T12:23:26.340Z",
      "content": "<p>During my EDA i came across few mask images which has no nonzero values, which means it has no useful information that could be of use . There are only 4 such mask which is not having any values </p>\n\n<p>out of this, <code>3790f55cad63053e956fb73027179707</code> is just a white image with no sample on it. </p>\n\n<p>&gt; 4a2ca53f240932e46eaf8959cb3f490a\n3790f55cad63053e956fb73027179707\ne4215cfc8c41ec04a55431cc413688a9\naaa5732cd49bffddf0d2b7d36fbb0a83</p>\n\n<p>Code to check\n```\nimport numpy as np\nimport skimage.io</p>\n\n<p>image_id = \"aaa5732cd49bffddf0d2b7d36fbb0a83\"</p>\n\n<p>img_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_images/\" + image_id + \".tiff\"\nmask_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_label_masks/\" + image_id \n+ \"_mask.tiff\"\nim = skimage.io.MultiImage(img_path)\nim_mask = skimage.io.MultiImage(mask_path)</p>\n\n<p>non_zero = np.nonzero(im_mask[-1])\nprint(\"img non zero values : \",non_zero)\n```</p>",
      "rawMarkdown": "During my EDA i came across few mask images which has no nonzero values, which means it has no useful information that could be of use . There are only 4 such mask which is not having any values \n\nout of this, `3790f55cad63053e956fb73027179707` is just a white image with no sample on it. \n\n\n&gt; 4a2ca53f240932e46eaf8959cb3f490a\n3790f55cad63053e956fb73027179707\ne4215cfc8c41ec04a55431cc413688a9\naaa5732cd49bffddf0d2b7d36fbb0a83\n\nCode to check\n```\nimport numpy as np\nimport skimage.io\n\nimage_id = \"aaa5732cd49bffddf0d2b7d36fbb0a83\"\n\nimg_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_images/\" + image_id + \".tiff\"\nmask_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_label_masks/\" + image_id \n+ \"_mask.tiff\"\nim = skimage.io.MultiImage(img_path)\nim_mask = skimage.io.MultiImage(mask_path)\n\nnon_zero = np.nonzero(im_mask[-1])\nprint(\"img non zero values : \",non_zero)\n```"
    }
  ],
  "comments": [
    {
      "id": 830982,
      "author_name": "Zac Dannelly",
      "author_url": "",
      "post_date": "2020-05-03T02:47:49.597000",
      "content": "<p>Added this note to <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/148060\">my thread tracking all odd slides</a>. If you come across more let me know and I can keep the running list updated!</p>",
      "votes": 1,
      "replies": [
        {
          "id": 831020,
          "author_name": "zhangeng",
          "author_url": "",
          "post_date": "2020-05-03T03:36:52.310000",
          "content": "<p><a href=\"https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data\">This is a list of more than 100 images waiting to be deleted</a></p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 831537,
          "author_name": "Zac Dannelly",
          "author_url": "",
          "post_date": "2020-05-03T12:25:55.650000",
          "content": "<p>Great thanks I added that to the discussion! </p>",
          "votes": 0,
          "replies": []
        },
        {
          "id": 832278,
          "author_name": "yuvaramsingh",
          "author_url": "",
          "post_date": "2020-05-04T02:40:05.950000",
          "content": "<p>sure  <a href=\"/dannellyz\">@dannellyz</a> </p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 829993,
      "author_name": "Benjamin Dubreu",
      "author_url": "",
      "post_date": "2020-05-02T08:31:23.780000",
      "content": "<p>Oh yeah, there is also a biopsy that is fully white. That the way it is with large datasets ;) </p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 829011,
      "author_name": "zhangeng",
      "author_url": "",
      "post_date": "2020-05-01T12:47:43.993000",
      "content": "<p>There is such  mask, which can be deleted or used as a verification set during training</p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "830982": "Added this note to [my thread tracking all odd slides](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/148060). If you come across more let me know and I can keep the running list updated!",
    "829993": "Oh yeah, there is also a biopsy that is fully white. That the way it is with large datasets ;) ",
    "829011": "There is such  mask, which can be deleted or used as a verification set during training",
    "828983": "During my EDA i came across few mask images which has no nonzero values, which means it has no useful information that could be of use . There are only 4 such mask which is not having any values \n\nout of this, `3790f55cad63053e956fb73027179707` is just a white image with no sample on it. \n\n\n&gt; 4a2ca53f240932e46eaf8959cb3f490a\n3790f55cad63053e956fb73027179707\ne4215cfc8c41ec04a55431cc413688a9\naaa5732cd49bffddf0d2b7d36fbb0a83\n\nCode to check\n```\nimport numpy as np\nimport skimage.io\n\nimage_id = \"aaa5732cd49bffddf0d2b7d36fbb0a83\"\n\nimg_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_images/\" + image_id + \".tiff\"\nmask_path = \"/kaggle/input/prostate-cancer-grade-assessment/train_label_masks/\" + image_id \n+ \"_mask.tiff\"\nim = skimage.io.MultiImage(img_path)\nim_mask = skimage.io.MultiImage(mask_path)\n\nnon_zero = np.nonzero(im_mask[-1])\nprint(\"img non zero values : \",non_zero)\n```"
  }
}