{
  "id": 417397,
  "title": "[72th solution]Simple bronze medal solution",
  "url": "/competitions/vesuvius-challenge-ink-detection/discussion/417397",
  "author_name": "Muj!rush!",
  "post_date": "2023-06-15T14:24:10.696000",
  "votes": 6,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Thank you for kaggle, host, and team KOKKO <a href=\"https://www.kaggle.com/coffeemountain\" target=\"_blank\">@coffeemountain</a> <a href=\"https://www.kaggle.com/hiroki1018\" target=\"_blank\">@hiroki1018</a> </p>\n<h1>Summary</h1>\n<ul>\n<li>simple post-processing in public notebook by using GaussianBlur.</li>\n<li>GaussianBlur parameter was decided public LB score.</li>\n<li>In this compe metrics, f0.5 is more importance precision than recall, and GaussianBlur is good effect for precision in this compe.</li>\n</ul>\n<h1>Code</h1>\n<ul>\n<li>We got bronze medal add below code in <a href=\"https://www.kaggle.com/code/yoyobar/3d-resnet-baseline-inference\" target=\"_blank\">this greate notebook</a>.</li>\n<li>Public LB 0.66, Private LB 0.59. </li>\n</ul>\n<pre><code>\n\nGaussianBlur_mask_pred = mask_pred\n\n kernel_size  (,, ): \n    GaussianBlur_mask_pred = cv2.GaussianBlur(GaussianBlur_mask_pred.astype(),(kernel_size,kernel_size),)\n\n\n</code></pre>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2F532522d88dd3b1b41c718790ce329b1f%2F2023-06-15%2022.52.10.png?generation=1686838525225224&amp;alt=media\" alt=\"\"></p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2Fdba5e0134037d793898550a3aa535c62%2F2023-06-15%2022.52.26.png?generation=1686838553940408&amp;alt=media\" alt=\"\"></p>\n<h1>Other</h1>\n<ul>\n<li>I could only do post-processing in this competition because I had no knowledge of segmentation tasks. I was lucky enough to get a bronze medal, but I would like to learn about the top solutions and aim for a higher rank next time!</li>\n</ul>",
  "messages": [
    {
      "id": 2303874,
      "postDate": "2023-06-15T14:24:10.697Z",
      "content": "<p>Thank you for kaggle, host, and team KOKKO <a href=\"https://www.kaggle.com/coffeemountain\" target=\"_blank\">@coffeemountain</a> <a href=\"https://www.kaggle.com/hiroki1018\" target=\"_blank\">@hiroki1018</a> </p>\n<h1>Summary</h1>\n<ul>\n<li>simple post-processing in public notebook by using GaussianBlur.</li>\n<li>GaussianBlur parameter was decided public LB score.</li>\n<li>In this compe metrics, f0.5 is more importance precision than recall, and GaussianBlur is good effect for precision in this compe.</li>\n</ul>\n<h1>Code</h1>\n<ul>\n<li>We got bronze medal add below code in <a href=\"https://www.kaggle.com/code/yoyobar/3d-resnet-baseline-inference\" target=\"_blank\">this greate notebook</a>.</li>\n<li>Public LB 0.66, Private LB 0.59. </li>\n</ul>\n<pre><code>\n\nGaussianBlur_mask_pred = mask_pred\n\n kernel_size  (,, ): \n    GaussianBlur_mask_pred = cv2.GaussianBlur(GaussianBlur_mask_pred.astype(),(kernel_size,kernel_size),)\n\n\n</code></pre>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2F532522d88dd3b1b41c718790ce329b1f%2F2023-06-15%2022.52.10.png?generation=1686838525225224&amp;alt=media\" alt=\"\"></p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2Fdba5e0134037d793898550a3aa535c62%2F2023-06-15%2022.52.26.png?generation=1686838553940408&amp;alt=media\" alt=\"\"></p>\n<h1>Other</h1>\n<ul>\n<li>I could only do post-processing in this competition because I had no knowledge of segmentation tasks. I was lucky enough to get a bronze medal, but I would like to learn about the top solutions and aim for a higher rank next time!</li>\n</ul>",
      "rawMarkdown": "Thank you for kaggle, host, and team KOKKO @coffeemountain @hiroki1018 \n\n# Summary\n* simple post-processing in public notebook by using GaussianBlur.\n* GaussianBlur parameter was decided public LB score.\n* In this compe metrics, f0.5 is more importance precision than recall, and GaussianBlur is good effect for precision in this compe.\n\n# Code\n* We got bronze medal add below code in [this greate notebook](https://www.kaggle.com/code/yoyobar/3d-resnet-baseline-inference).\n* Public LB 0.66, Private LB 0.59. \n\n```python\n################ GaussianBlur ################ \n\nGaussianBlur_mask_pred = mask_pred\n\nfor kernel_size in range(1,94, 2): \n    GaussianBlur_mask_pred = cv2.GaussianBlur(GaussianBlur_mask_pred.astype('uint8'),(kernel_size,kernel_size),0)\n\n############################################\n```\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2F532522d88dd3b1b41c718790ce329b1f%2F2023-06-15%2022.52.10.png?generation=1686838525225224&alt=media)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2Fdba5e0134037d793898550a3aa535c62%2F2023-06-15%2022.52.26.png?generation=1686838553940408&alt=media)\n\n\n# Other\n* I could only do post-processing in this competition because I had no knowledge of segmentation tasks. I was lucky enough to get a bronze medal, but I would like to learn about the top solutions and aim for a higher rank next time!",
      "votes": 6
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2303874": "Thank you for kaggle, host, and team KOKKO @coffeemountain @hiroki1018 \n\n# Summary\n* simple post-processing in public notebook by using GaussianBlur.\n* GaussianBlur parameter was decided public LB score.\n* In this compe metrics, f0.5 is more importance precision than recall, and GaussianBlur is good effect for precision in this compe.\n\n# Code\n* We got bronze medal add below code in [this greate notebook](https://www.kaggle.com/code/yoyobar/3d-resnet-baseline-inference).\n* Public LB 0.66, Private LB 0.59. \n\n```python\n################ GaussianBlur ################ \n\nGaussianBlur_mask_pred = mask_pred\n\nfor kernel_size in range(1,94, 2): \n    GaussianBlur_mask_pred = cv2.GaussianBlur(GaussianBlur_mask_pred.astype('uint8'),(kernel_size,kernel_size),0)\n\n############################################\n```\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2F532522d88dd3b1b41c718790ce329b1f%2F2023-06-15%2022.52.10.png?generation=1686838525225224&alt=media)\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F7082245%2Fdba5e0134037d793898550a3aa535c62%2F2023-06-15%2022.52.26.png?generation=1686838553940408&alt=media)\n\n\n# Other\n* I could only do post-processing in this competition because I had no knowledge of segmentation tasks. I was lucky enough to get a bronze medal, but I would like to learn about the top solutions and aim for a higher rank next time!"
  }
}