{
  "id": 151098,
  "title": "PANDA level 1 multipart dataset available",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/151098",
  "author_name": "Yovin Yahathugoda",
  "post_date": "2020-05-14T10:44:20.240000",
  "votes": 6,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Since everyone does not have the resources to download the entire dataset at once i have created a 3 part dataset of level 1 training images. I have not cropped, rescaled or resized any of the images. Only the training images are available, no masks have been included. Hope some of you find the dataset useful.</p>\n\n<p><a href=\"https://www.kaggle.com/yovinyahathugoda/panda-level1-part1\">https://www.kaggle.com/yovinyahathugoda/panda-level1-part1</a>\n<a href=\"https://www.kaggle.com/yovinyahathugoda/panda-level1-part2\">https://www.kaggle.com/yovinyahathugoda/panda-level1-part2</a>\n<a href=\"https://www.kaggle.com/yovinyahathugoda/panda-level1-part3\">https://www.kaggle.com/yovinyahathugoda/panda-level1-part3</a></p>",
  "messages": [
    {
      "id": 847360,
      "postDate": "2020-05-14T10:44:20.240Z",
      "content": "<p>Since everyone does not have the resources to download the entire dataset at once i have created a 3 part dataset of level 1 training images. I have not cropped, rescaled or resized any of the images. Only the training images are available, no masks have been included. Hope some of you find the dataset useful.</p>\n\n<p><a href=\"https://www.kaggle.com/yovinyahathugoda/panda-level1-part1\">https://www.kaggle.com/yovinyahathugoda/panda-level1-part1</a>\n<a href=\"https://www.kaggle.com/yovinyahathugoda/panda-level1-part2\">https://www.kaggle.com/yovinyahathugoda/panda-level1-part2</a>\n<a href=\"https://www.kaggle.com/yovinyahathugoda/panda-level1-part3\">https://www.kaggle.com/yovinyahathugoda/panda-level1-part3</a></p>",
      "rawMarkdown": "Since everyone does not have the resources to download the entire dataset at once i have created a 3 part dataset of level 1 training images. I have not cropped, rescaled or resized any of the images. Only the training images are available, no masks have been included. Hope some of you find the dataset useful.\n\n[https://www.kaggle.com/yovinyahathugoda/panda-level1-part1](https://www.kaggle.com/yovinyahathugoda/panda-level1-part1)\n[https://www.kaggle.com/yovinyahathugoda/panda-level1-part2](https://www.kaggle.com/yovinyahathugoda/panda-level1-part2)\n[https://www.kaggle.com/yovinyahathugoda/panda-level1-part3](https://www.kaggle.com/yovinyahathugoda/panda-level1-part3)",
      "votes": 6
    },
    {
      "id": 883300,
      "postDate": "2020-06-12T14:20:18.197Z",
      "content": "<p><a href=\"/micheomaano\">@micheomaano</a> attached is the code i used to generate the files. All i did was load the level 1 image and save it directly as png.</p>\n\n<p>```\nall_image_ids = list(train_data_df.image_id)</p>\n\n<p>def make_images(image_id):\n    load_path = slide_dir + image_id + '.tiff'\n    save_path = save_dir + image_id + '.png'</p>\n\n<pre><code>slide = skimage.io.MultiImage(load_path)[1]\n\ncv2.imwrite(save_path, slide)\nreturn image_id\n</code></pre>\n\n<p>with Pool(processes=4) as pool:\n    show_run_results = list(\n        tqdm(pool.imap(make_images, all_image_ids), total = len(all_image_ids ))\n    )\n```</p>",
      "rawMarkdown": "@micheomaano attached is the code i used to generate the files. All i did was load the level 1 image and save it directly as png.\n\n```\nall_image_ids = list(train_data_df.image_id)\n\ndef make_images(image_id):\n    load_path = slide_dir + image_id + '.tiff'\n    save_path = save_dir + image_id + '.png'\n    \n    slide = skimage.io.MultiImage(load_path)[1]\n\n    cv2.imwrite(save_path, slide)\n    return image_id\n        \n\nwith Pool(processes=4) as pool:\n    show_run_results = list(\n        tqdm(pool.imap(make_images, all_image_ids), total = len(all_image_ids ))\n    )\n```",
      "replies": [
        {
          "id": 883304,
          "postDate": "2020-06-12T14:22:06.593Z",
          "content": "<p>Thanks.</p>",
          "rawMarkdown": "Thanks.\n"
        }
      ]
    },
    {
      "id": 883159,
      "postDate": "2020-06-12T12:13:34.217Z",
      "content": "<p>Can you please tell me how you generated this data?\nLike its code so that I can view its authenticity or something like that. \nI really need a Dataset for level 1 images.\nThanks btw. 😊</p>",
      "rawMarkdown": "Can you please tell me how you generated this data?\nLike its code so that I can view its authenticity or something like that. \nI really need a Dataset for level 1 images.\nThanks btw. 😊"
    }
  ],
  "comments": [
    {
      "id": 883300,
      "author_name": "Yovin Yahathugoda",
      "author_url": "",
      "post_date": "2020-06-12T14:20:18.197000",
      "content": "<p><a href=\"/micheomaano\">@micheomaano</a> attached is the code i used to generate the files. All i did was load the level 1 image and save it directly as png.</p>\n\n<p>```\nall_image_ids = list(train_data_df.image_id)</p>\n\n<p>def make_images(image_id):\n    load_path = slide_dir + image_id + '.tiff'\n    save_path = save_dir + image_id + '.png'</p>\n\n<pre><code>slide = skimage.io.MultiImage(load_path)[1]\n\ncv2.imwrite(save_path, slide)\nreturn image_id\n</code></pre>\n\n<p>with Pool(processes=4) as pool:\n    show_run_results = list(\n        tqdm(pool.imap(make_images, all_image_ids), total = len(all_image_ids ))\n    )\n```</p>",
      "votes": 0,
      "replies": [
        {
          "id": 883304,
          "author_name": "Salman",
          "author_url": "",
          "post_date": "2020-06-12T14:22:06.593000",
          "content": "<p>Thanks.</p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 883159,
      "author_name": "Salman",
      "author_url": "",
      "post_date": "2020-06-12T12:13:34.217000",
      "content": "<p>Can you please tell me how you generated this data?\nLike its code so that I can view its authenticity or something like that. \nI really need a Dataset for level 1 images.\nThanks btw. 😊</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "847360": "Since everyone does not have the resources to download the entire dataset at once i have created a 3 part dataset of level 1 training images. I have not cropped, rescaled or resized any of the images. Only the training images are available, no masks have been included. Hope some of you find the dataset useful.\n\n[https://www.kaggle.com/yovinyahathugoda/panda-level1-part1](https://www.kaggle.com/yovinyahathugoda/panda-level1-part1)\n[https://www.kaggle.com/yovinyahathugoda/panda-level1-part2](https://www.kaggle.com/yovinyahathugoda/panda-level1-part2)\n[https://www.kaggle.com/yovinyahathugoda/panda-level1-part3](https://www.kaggle.com/yovinyahathugoda/panda-level1-part3)",
    "883300": "@micheomaano attached is the code i used to generate the files. All i did was load the level 1 image and save it directly as png.\n\n```\nall_image_ids = list(train_data_df.image_id)\n\ndef make_images(image_id):\n    load_path = slide_dir + image_id + '.tiff'\n    save_path = save_dir + image_id + '.png'\n    \n    slide = skimage.io.MultiImage(load_path)[1]\n\n    cv2.imwrite(save_path, slide)\n    return image_id\n        \n\nwith Pool(processes=4) as pool:\n    show_run_results = list(\n        tqdm(pool.imap(make_images, all_image_ids), total = len(all_image_ids ))\n    )\n```",
    "883159": "Can you please tell me how you generated this data?\nLike its code so that I can view its authenticity or something like that. \nI really need a Dataset for level 1 images.\nThanks btw. 😊"
  }
}