{
  "id": 148060,
  "title": "Blank / Error / Marking / Odd Slide Thread",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/148060",
  "author_name": "Zac Dannelly",
  "post_date": "2020-05-03T02:38:31.385000",
  "votes": 27,
  "comment_count": 7,
  "views": 0,
  "content": "<h3>EDIT MAY14</h3>\n\n<p>I made another <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/151323\">discussion post </a> that outlined my <a href=\"https://www.kaggle.com/dannellyz/collection-of-600-suspicious-slides-data-loader/\">notebook</a> putting this info into practise. </p>\n\n<p>I wanted to get a thread started that compiled a list of <code>image_ids</code> that have something odd about them. I know that as I have been doing by analysis I have come across a few that have messed up an otherwise nice pipeline. With more that 10k images I rather just drop those few images that cause me large amounts of try/except frameworks. I also think the test images will have been scraped for such errors. I have pulled some from other discussion threads, but please comment those you find and I can update the list.</p>\n\n<h1>Blank</h1>\n\n<p><code>\nblank_slides = [\"3790f55cad63053e956fb73027179707\"]\nblank_masks = [\"4a2ca53f240932e46eaf8959cb3f490a\", \"3790f55cad63053e956fb73027179707\", \n\"e4215cfc8c41ec04a55431cc413688a9\", \n\"aaa5732cd49bffddf0d2b7d36fbb0a83\"]\n</code></p>\n\n<h2>Contributors:</h2>\n\n<ul>\n<li>@yuvaramsing - <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/147688\">Post</a></li>\n</ul>\n\n<h1>Low Tissue Count (&lt; 3% of slide)</h1>\n\n<p><code>\nlow_tiss_slides = ['033e39459301e97e457232780a314ab7',\n '0b6e34bf65ee0810c1a4bf702b667c88',\n '3385a0f7f4f3e7e7b380325582b115c9',\n '3790f55cad63053e956fb73027179707',\n '5204134e82ce75b1109cc1913d81abc6',\n 'a08e24cff451d628df797efc4343e13c']\n</code></p>\n\n<h1>Pen Markings</h1>\n\n<ul>\n<li><a href=\"https://www.kaggle.com/rohitsingh9990/marker-images\">Dataset</a> by <a href=\"/rohitsingh9990\">@rohitsingh9990</a> with comprehensive list. Great <a href=\"https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline\">EDA Notebook</a> as well to discover these items. </li>\n</ul>\n\n<h1>Suspicious Masks</h1>\n\n<ul>\n<li><a href=\"/yuvaramsingh\">@yuvaramsingh</a> made me aware of the very nice EDA done by <a href=\"/iamleonie\">@iamleonie</a> in the notebook: <a href=\"https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data/output\">PANDA: EDA, Visualizations &amp; Suspicious Data</a>. As a part of the output in that notebook there is a great csv that outlines more slides with suspicious features.</li>\n</ul>",
  "messages": [
    {
      "id": 830970,
      "postDate": "2020-05-03T02:38:31.387Z",
      "content": "<h3>EDIT MAY14</h3>\n\n<p>I made another <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/151323\">discussion post </a> that outlined my <a href=\"https://www.kaggle.com/dannellyz/collection-of-600-suspicious-slides-data-loader/\">notebook</a> putting this info into practise. </p>\n\n<p>I wanted to get a thread started that compiled a list of <code>image_ids</code> that have something odd about them. I know that as I have been doing by analysis I have come across a few that have messed up an otherwise nice pipeline. With more that 10k images I rather just drop those few images that cause me large amounts of try/except frameworks. I also think the test images will have been scraped for such errors. I have pulled some from other discussion threads, but please comment those you find and I can update the list.</p>\n\n<h1>Blank</h1>\n\n<p><code>\nblank_slides = [\"3790f55cad63053e956fb73027179707\"]\nblank_masks = [\"4a2ca53f240932e46eaf8959cb3f490a\", \"3790f55cad63053e956fb73027179707\", \n\"e4215cfc8c41ec04a55431cc413688a9\", \n\"aaa5732cd49bffddf0d2b7d36fbb0a83\"]\n</code></p>\n\n<h2>Contributors:</h2>\n\n<ul>\n<li>@yuvaramsing - <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/147688\">Post</a></li>\n</ul>\n\n<h1>Low Tissue Count (&lt; 3% of slide)</h1>\n\n<p><code>\nlow_tiss_slides = ['033e39459301e97e457232780a314ab7',\n '0b6e34bf65ee0810c1a4bf702b667c88',\n '3385a0f7f4f3e7e7b380325582b115c9',\n '3790f55cad63053e956fb73027179707',\n '5204134e82ce75b1109cc1913d81abc6',\n 'a08e24cff451d628df797efc4343e13c']\n</code></p>\n\n<h1>Pen Markings</h1>\n\n<ul>\n<li><a href=\"https://www.kaggle.com/rohitsingh9990/marker-images\">Dataset</a> by <a href=\"/rohitsingh9990\">@rohitsingh9990</a> with comprehensive list. Great <a href=\"https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline\">EDA Notebook</a> as well to discover these items. </li>\n</ul>\n\n<h1>Suspicious Masks</h1>\n\n<ul>\n<li><a href=\"/yuvaramsingh\">@yuvaramsingh</a> made me aware of the very nice EDA done by <a href=\"/iamleonie\">@iamleonie</a> in the notebook: <a href=\"https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data/output\">PANDA: EDA, Visualizations &amp; Suspicious Data</a>. As a part of the output in that notebook there is a great csv that outlines more slides with suspicious features.</li>\n</ul>",
      "rawMarkdown": "### EDIT MAY14\nI made another [discussion post ](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/151323) that outlined my [notebook](https://www.kaggle.com/dannellyz/collection-of-600-suspicious-slides-data-loader/) putting this info into practise. \n\nI wanted to get a thread started that compiled a list of `image_ids` that have something odd about them. I know that as I have been doing by analysis I have come across a few that have messed up an otherwise nice pipeline. With more that 10k images I rather just drop those few images that cause me large amounts of try/except frameworks. I also think the test images will have been scraped for such errors. I have pulled some from other discussion threads, but please comment those you find and I can update the list.\n\n# Blank\n```\nblank_slides = [\"3790f55cad63053e956fb73027179707\"]\nblank_masks = [\"4a2ca53f240932e46eaf8959cb3f490a\", \"3790f55cad63053e956fb73027179707\", \n\"e4215cfc8c41ec04a55431cc413688a9\", \n\"aaa5732cd49bffddf0d2b7d36fbb0a83\"]\n```\n## Contributors:\n- @yuvaramsing - [Post](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/147688)\n\n# Low Tissue Count (&lt; 3% of slide)\n```\nlow_tiss_slides = ['033e39459301e97e457232780a314ab7',\n '0b6e34bf65ee0810c1a4bf702b667c88',\n '3385a0f7f4f3e7e7b380325582b115c9',\n '3790f55cad63053e956fb73027179707',\n '5204134e82ce75b1109cc1913d81abc6',\n 'a08e24cff451d628df797efc4343e13c']\n```\n\n#Pen Markings\n- [Dataset](https://www.kaggle.com/rohitsingh9990/marker-images) by @rohitsingh9990 with comprehensive list. Great [EDA Notebook](https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline) as well to discover these items. \n\n#Suspicious Masks\n- @yuvaramsingh made me aware of the very nice EDA done by @iamleonie in the notebook: [PANDA: EDA, Visualizations &amp; Suspicious Data](https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data/output). As a part of the output in that notebook there is a great csv that outlines more slides with suspicious features.\n",
      "votes": 27
    },
    {
      "id": 831932,
      "postDate": "2020-05-03T17:33:20.730Z",
      "content": "<p>If any one interested in the pen marked images dataset, here i have created one <a href=\"https://www.kaggle.com/rohitsingh9990/marker-images\">https://www.kaggle.com/rohitsingh9990/marker-images</a></p>",
      "rawMarkdown": "If any one interested in the pen marked images dataset, here i have created one https://www.kaggle.com/rohitsingh9990/marker-images",
      "votes": 6,
      "replies": [
        {
          "id": 832022,
          "postDate": "2020-05-03T18:31:01.840Z",
          "content": "<p>Awesome added to the post thanks! Thinking about consolidating the findings eventually into a dataset on Kaggle as well. Would it be okay to include these?</p>",
          "rawMarkdown": "Awesome added to the post thanks! Thinking about consolidating the findings eventually into a dataset on Kaggle as well. Would it be okay to include these?",
          "votes": 1
        },
        {
          "id": 832024,
          "postDate": "2020-05-03T18:39:02.247Z",
          "content": "<p>yes, its fine.</p>",
          "rawMarkdown": "yes, its fine.",
          "votes": 1
        },
        {
          "id": 849339,
          "postDate": "2020-05-15T17:10:55.957Z",
          "content": "<p><a href=\"https://www.kaggle.com/akensert/panda-removal-of-pen-marks\">Here</a>'s my attempt to remove pen marks from (some) the images!</p>",
          "rawMarkdown": "[Here](https://www.kaggle.com/akensert/panda-removal-of-pen-marks)'s my attempt to remove pen marks from (some) the images!"
        }
      ]
    },
    {
      "id": 831925,
      "postDate": "2020-05-03T17:28:39.093Z",
      "content": "<p>Great <a href=\"/dannellyz\">@dannellyz</a>! This kernel by <a href=\"/rohitsingh9990\">@rohitsingh9990</a> has a list of penmarks, may be you can add to the post? <a href=\"https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline\">https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline</a> </p>",
      "rawMarkdown": "Great @dannellyz! This kernel by @rohitsingh9990 has a list of penmarks, may be you can add to the post? https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline ",
      "votes": 2,
      "replies": [
        {
          "id": 832023,
          "postDate": "2020-05-03T18:31:16.267Z",
          "content": "<p>Thanks for the note added!!</p>",
          "rawMarkdown": "Thanks for the note added!!",
          "votes": 1
        }
      ]
    },
    {
      "id": 834663,
      "postDate": "2020-05-05T17:21:42.543Z",
      "content": "<p>Thanks for the note</p>",
      "rawMarkdown": "Thanks for the note"
    }
  ],
  "comments": [
    {
      "id": 831932,
      "author_name": "Dracarys",
      "author_url": "",
      "post_date": "2020-05-03T17:33:20.730000",
      "content": "<p>If any one interested in the pen marked images dataset, here i have created one <a href=\"https://www.kaggle.com/rohitsingh9990/marker-images\">https://www.kaggle.com/rohitsingh9990/marker-images</a></p>",
      "votes": 6,
      "replies": [
        {
          "id": 832022,
          "author_name": "Zac Dannelly",
          "author_url": "",
          "post_date": "2020-05-03T18:31:01.840000",
          "content": "<p>Awesome added to the post thanks! Thinking about consolidating the findings eventually into a dataset on Kaggle as well. Would it be okay to include these?</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 832024,
          "author_name": "Dracarys",
          "author_url": "",
          "post_date": "2020-05-03T18:39:02.247000",
          "content": "<p>yes, its fine.</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 849339,
          "author_name": "Alex",
          "author_url": "",
          "post_date": "2020-05-15T17:10:55.957000",
          "content": "<p><a href=\"https://www.kaggle.com/akensert/panda-removal-of-pen-marks\">Here</a>'s my attempt to remove pen marks from (some) the images!</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 831925,
      "author_name": "Debanga Raj Neog",
      "author_url": "",
      "post_date": "2020-05-03T17:28:39.093000",
      "content": "<p>Great <a href=\"/dannellyz\">@dannellyz</a>! This kernel by <a href=\"/rohitsingh9990\">@rohitsingh9990</a> has a list of penmarks, may be you can add to the post? <a href=\"https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline\">https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline</a> </p>",
      "votes": 2,
      "replies": [
        {
          "id": 832023,
          "author_name": "Zac Dannelly",
          "author_url": "",
          "post_date": "2020-05-03T18:31:16.267000",
          "content": "<p>Thanks for the note added!!</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 834663,
      "author_name": "Subhashini Mariappan",
      "author_url": "",
      "post_date": "2020-05-05T17:21:42.543000",
      "content": "<p>Thanks for the note</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "830970": "### EDIT MAY14\nI made another [discussion post ](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/151323) that outlined my [notebook](https://www.kaggle.com/dannellyz/collection-of-600-suspicious-slides-data-loader/) putting this info into practise. \n\nI wanted to get a thread started that compiled a list of `image_ids` that have something odd about them. I know that as I have been doing by analysis I have come across a few that have messed up an otherwise nice pipeline. With more that 10k images I rather just drop those few images that cause me large amounts of try/except frameworks. I also think the test images will have been scraped for such errors. I have pulled some from other discussion threads, but please comment those you find and I can update the list.\n\n# Blank\n```\nblank_slides = [\"3790f55cad63053e956fb73027179707\"]\nblank_masks = [\"4a2ca53f240932e46eaf8959cb3f490a\", \"3790f55cad63053e956fb73027179707\", \n\"e4215cfc8c41ec04a55431cc413688a9\", \n\"aaa5732cd49bffddf0d2b7d36fbb0a83\"]\n```\n## Contributors:\n- @yuvaramsing - [Post](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/147688)\n\n# Low Tissue Count (&lt; 3% of slide)\n```\nlow_tiss_slides = ['033e39459301e97e457232780a314ab7',\n '0b6e34bf65ee0810c1a4bf702b667c88',\n '3385a0f7f4f3e7e7b380325582b115c9',\n '3790f55cad63053e956fb73027179707',\n '5204134e82ce75b1109cc1913d81abc6',\n 'a08e24cff451d628df797efc4343e13c']\n```\n\n#Pen Markings\n- [Dataset](https://www.kaggle.com/rohitsingh9990/marker-images) by @rohitsingh9990 with comprehensive list. Great [EDA Notebook](https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline) as well to discover these items. \n\n#Suspicious Masks\n- @yuvaramsingh made me aware of the very nice EDA done by @iamleonie in the notebook: [PANDA: EDA, Visualizations &amp; Suspicious Data](https://www.kaggle.com/iamleonie/panda-eda-visualizations-suspicious-data/output). As a part of the output in that notebook there is a great csv that outlines more slides with suspicious features.\n",
    "831932": "If any one interested in the pen marked images dataset, here i have created one https://www.kaggle.com/rohitsingh9990/marker-images",
    "831925": "Great @dannellyz! This kernel by @rohitsingh9990 has a list of penmarks, may be you can add to the post? https://www.kaggle.com/rohitsingh9990/panda-eda-better-visualization-simple-baseline ",
    "834663": "Thanks for the note"
  }
}