{
  "id": 216430,
  "title": "Train Dataset Containing All Cell Masks as RLE Strings",
  "url": "/competitions/hpa-single-cell-image-classification/discussion/216430",
  "author_name": "Darien Schettler",
  "post_date": "2021-02-02T16:59:27.058000",
  "votes": 13,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi there. I hope you're all doing well!</p>\n<p>I created a CSV containing all cell-wise segmentation masks represented as strings after Run-Length-Encoding. </p>\n<p>The cell-wise segmentation masks originate from the user Titos <a href=\"https://www.kaggle.com/c/hpa-single-cell-image-classification/discussion/215773\" target=\"_blank\">dataset</a> dataset that was generated using the CellSegmentator tool. </p>\n<p>You can find the CSV file here  --&gt; <a href=\"https://www.kaggle.com/dschettler8845/hpa-processed-train-dataframe-with-cellwise-rle\" target=\"_blank\">https://www.kaggle.com/dschettler8845/hpa-processed-train-dataframe-with-cellwise-rle</a></p>\n<p>I also created a <a href=\"https://www.kaggle.com/dschettler8845/starter-on-how-to-load-rle-masks\" target=\"_blank\">notebook</a> to show how to open and process the file (30-60 seconds) as the list of rle's will be interpreted as one big string (contains decode/encode/preprocess info).</p>\n<p>Hope this helps.</p>",
  "messages": [
    {
      "id": 1182968,
      "postDate": "2021-02-02T16:59:27.060Z",
      "content": "<p>Hi there. I hope you're all doing well!</p>\n<p>I created a CSV containing all cell-wise segmentation masks represented as strings after Run-Length-Encoding. </p>\n<p>The cell-wise segmentation masks originate from the user Titos <a href=\"https://www.kaggle.com/c/hpa-single-cell-image-classification/discussion/215773\" target=\"_blank\">dataset</a> dataset that was generated using the CellSegmentator tool. </p>\n<p>You can find the CSV file here  --&gt; <a href=\"https://www.kaggle.com/dschettler8845/hpa-processed-train-dataframe-with-cellwise-rle\" target=\"_blank\">https://www.kaggle.com/dschettler8845/hpa-processed-train-dataframe-with-cellwise-rle</a></p>\n<p>I also created a <a href=\"https://www.kaggle.com/dschettler8845/starter-on-how-to-load-rle-masks\" target=\"_blank\">notebook</a> to show how to open and process the file (30-60 seconds) as the list of rle's will be interpreted as one big string (contains decode/encode/preprocess info).</p>\n<p>Hope this helps.</p>",
      "rawMarkdown": "Hi there. I hope you're all doing well!\n\nI created a CSV containing all cell-wise segmentation masks represented as strings after Run-Length-Encoding. \n\nThe cell-wise segmentation masks originate from the user Titos [dataset](https://www.kaggle.com/c/hpa-single-cell-image-classification/discussion/215773) dataset that was generated using the CellSegmentator tool. \n\nYou can find the CSV file here  --> https://www.kaggle.com/dschettler8845/hpa-processed-train-dataframe-with-cellwise-rle\n\nI also created a [notebook](https://www.kaggle.com/dschettler8845/starter-on-how-to-load-rle-masks) to show how to open and process the file (30-60 seconds) as the list of rle's will be interpreted as one big string (contains decode/encode/preprocess info).\n\nHope this helps.",
      "votes": 13
    },
    {
      "id": 1213086,
      "postDate": "2021-02-21T20:36:25.567Z",
      "content": "<p>Thanks for sharing! You are a true hero.<br>\nJust wondering if you can add masks of nuclei to your csv as well? Titos dataset includes those too. I can help if you don't have the time. Just don't want to create a separate dataset which contains the same information.</p>",
      "rawMarkdown": "Thanks for sharing! You are a true hero.\nJust wondering if you can add masks of nuclei to your csv as well? Titos dataset includes those too. I can help if you don't have the time. Just don't want to create a separate dataset which contains the same information.",
      "votes": 1,
      "replies": [
        {
          "id": 1213154,
          "postDate": "2021-02-21T22:20:54.870Z",
          "content": "<p>I will update it and comment here when it's done! Thanks for the feedback!</p>",
          "rawMarkdown": "I will update it and comment here when it's done! Thanks for the feedback!",
          "votes": 1
        },
        {
          "id": 1220914,
          "postDate": "2021-02-28T13:31:13.960Z",
          "content": "<p><a href=\"https://www.kaggle.com/jihangz\" target=\"_blank\">@jihangz</a> - Apologies, I just reran this whole process (from raw not from Tito's) and I forgot to save the nuclei masks.</p>\n<p>I don't plan on rerunning it again anytime soon. If you do plan to run it yourself, I can make you a contributor to the dataset and you can update it when you're ready?</p>",
          "rawMarkdown": "@jihangz - Apologies, I just reran this whole process (from raw not from Tito's) and I forgot to save the nuclei masks.\n\nI don't plan on rerunning it again anytime soon. If you do plan to run it yourself, I can make you a contributor to the dataset and you can update it when you're ready?"
        }
      ]
    },
    {
      "id": 1193846,
      "postDate": "2021-02-09T22:20:43.753Z",
      "content": "<p>Just wanted to update. This dataset now includes the bounding-boxes for each RLE mask.</p>\n<p>So each row in the dataframe is an image the <strong><code>cell_masks</code></strong> column contains a list of masks for every cell in the respective image. Now we also have the <strong><code>bboxes</code></strong> column which contains a list of bounding-boxes (<strong><code>xmin, ymin, xmax, ymax</code></strong>) for every respective cell in the respective image. </p>",
      "rawMarkdown": "Just wanted to update. This dataset now includes the bounding-boxes for each RLE mask.\n\nSo each row in the dataframe is an image the **`cell_masks`** column contains a list of masks for every cell in the respective image. Now we also have the **`bboxes`** column which contains a list of bounding-boxes (**`xmin, ymin, xmax, ymax`**) for every respective cell in the respective image. ",
      "votes": 2
    }
  ],
  "comments": [
    {
      "id": 1213086,
      "author_name": "Darkate",
      "author_url": "",
      "post_date": "2021-02-21T20:36:25.567000",
      "content": "<p>Thanks for sharing! You are a true hero.<br>\nJust wondering if you can add masks of nuclei to your csv as well? Titos dataset includes those too. I can help if you don't have the time. Just don't want to create a separate dataset which contains the same information.</p>",
      "votes": 1,
      "replies": [
        {
          "id": 1213154,
          "author_name": "Darien Schettler",
          "author_url": "",
          "post_date": "2021-02-21T22:20:54.870000",
          "content": "<p>I will update it and comment here when it's done! Thanks for the feedback!</p>",
          "votes": 1,
          "replies": []
        },
        {
          "id": 1220914,
          "author_name": "Darien Schettler",
          "author_url": "",
          "post_date": "2021-02-28T13:31:13.960000",
          "content": "<p><a href=\"https://www.kaggle.com/jihangz\" target=\"_blank\">@jihangz</a> - Apologies, I just reran this whole process (from raw not from Tito's) and I forgot to save the nuclei masks.</p>\n<p>I don't plan on rerunning it again anytime soon. If you do plan to run it yourself, I can make you a contributor to the dataset and you can update it when you're ready?</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 1193846,
      "author_name": "Darien Schettler",
      "author_url": "",
      "post_date": "2021-02-09T22:20:43.753000",
      "content": "<p>Just wanted to update. This dataset now includes the bounding-boxes for each RLE mask.</p>\n<p>So each row in the dataframe is an image the <strong><code>cell_masks</code></strong> column contains a list of masks for every cell in the respective image. Now we also have the <strong><code>bboxes</code></strong> column which contains a list of bounding-boxes (<strong><code>xmin, ymin, xmax, ymax</code></strong>) for every respective cell in the respective image. </p>",
      "votes": 2,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1182968": "Hi there. I hope you're all doing well!\n\nI created a CSV containing all cell-wise segmentation masks represented as strings after Run-Length-Encoding. \n\nThe cell-wise segmentation masks originate from the user Titos [dataset](https://www.kaggle.com/c/hpa-single-cell-image-classification/discussion/215773) dataset that was generated using the CellSegmentator tool. \n\nYou can find the CSV file here  --> https://www.kaggle.com/dschettler8845/hpa-processed-train-dataframe-with-cellwise-rle\n\nI also created a [notebook](https://www.kaggle.com/dschettler8845/starter-on-how-to-load-rle-masks) to show how to open and process the file (30-60 seconds) as the list of rle's will be interpreted as one big string (contains decode/encode/preprocess info).\n\nHope this helps.",
    "1213086": "Thanks for sharing! You are a true hero.\nJust wondering if you can add masks of nuclei to your csv as well? Titos dataset includes those too. I can help if you don't have the time. Just don't want to create a separate dataset which contains the same information.",
    "1193846": "Just wanted to update. This dataset now includes the bounding-boxes for each RLE mask.\n\nSo each row in the dataframe is an image the **`cell_masks`** column contains a list of masks for every cell in the respective image. Now we also have the **`bboxes`** column which contains a list of bounding-boxes (**`xmin, ymin, xmax, ymax`**) for every respective cell in the respective image. "
  }
}