{
  "id": 252602,
  "title": "Cropping CXRs with YOLOv5",
  "url": "/competitions/siim-covid19-detection/discussion/252602",
  "author_name": "",
  "post_date": "2021-07-13T04:39:25.220878100Z",
  "votes": 4,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I made a Torch/YOLOv5 model to detect lungs, clavicles, shoulders and the heart. The idea is to figure out where to crop in order to eliminate as much 'noise' as possible.</p>\n<p>I call it 'Smart Cropping'</p>\n<p>This notebook demonstrates demonstrates predicting bounding boxes on anatomy, and then cropping the image to fit.</p>\n<p><a href=\"https://www.kaggle.com/davidbroberts/smart-cropping-cxr-with-yolov5\" target=\"_blank\">https://www.kaggle.com/davidbroberts/smart-cropping-cxr-with-yolov5</a></p>",
  "messages": [
    {
      "id": "1385853",
      "postDate": "07/13/2021 04:39:25",
      "content": "<p>I made a Torch/YOLOv5 model to detect lungs, clavicles, shoulders and the heart. The idea is to figure out where to crop in order to eliminate as much 'noise' as possible.</p>\n<p>I call it 'Smart Cropping'</p>\n<p>This notebook demonstrates demonstrates predicting bounding boxes on anatomy, and then cropping the image to fit.</p>\n<p><a href=\"https://www.kaggle.com/davidbroberts/smart-cropping-cxr-with-yolov5\" target=\"_blank\">https://www.kaggle.com/davidbroberts/smart-cropping-cxr-with-yolov5</a></p>",
      "rawMarkdown": "I made a Torch/YOLOv5 model to detect lungs, clavicles, shoulders and the heart. The idea is to figure out where to crop in order to eliminate as much 'noise' as possible.\n\nI call it 'Smart Cropping'\n\nThis notebook demonstrates demonstrates predicting bounding boxes on anatomy, and then cropping the image to fit.\n\nhttps://www.kaggle.com/davidbroberts/smart-cropping-cxr-with-yolov5",
      "votes": null
    },
    {
      "id": "1386702",
      "postDate": "07/13/2021 16:24:54",
      "content": "<p>I updated the anatomy detection model in this notebook to include the trachea, carina and gastric bubble.</p>",
      "rawMarkdown": "I updated the anatomy detection model in this notebook to include the trachea, carina and gastric bubble.",
      "votes": null
    },
    {
      "id": "1391435",
      "postDate": "07/17/2021 15:40:00",
      "content": "<p>Good work!</p>",
      "rawMarkdown": "Good work!",
      "votes": null
    },
    {
      "id": "1396880",
      "postDate": "07/22/2021 14:58:56",
      "content": "<p>I made a dataset and a train notebook for base anatomy detection model.</p>\n<p>Dataset -&gt; <a href=\"https://www.kaggle.com/davidbroberts/chest-anatomy-yolov5-train-set\" target=\"_blank\">https://www.kaggle.com/davidbroberts/chest-anatomy-yolov5-train-set</a><br>\nTrain notebook -&gt; <a href=\"https://www.kaggle.com/davidbroberts/chest-anatomy-detection-train\" target=\"_blank\">https://www.kaggle.com/davidbroberts/chest-anatomy-detection-train</a></p>\n<p>Feel free to use it as you wish. The images are from the SIIM-COVID19 dataset, so the original license applies.</p>",
      "rawMarkdown": "I made a dataset and a train notebook for base anatomy detection model.\n\nDataset -> https://www.kaggle.com/davidbroberts/chest-anatomy-yolov5-train-set\nTrain notebook -> https://www.kaggle.com/davidbroberts/chest-anatomy-detection-train\n\nFeel free to use it as you wish. The images are from the SIIM-COVID19 dataset, so the original license applies.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1386702,
      "author_name": "davidbroberts",
      "author_url": "",
      "post_date": "07/13/2021 16:24:54",
      "content": "<p>I updated the anatomy detection model in this notebook to include the trachea, carina and gastric bubble.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1391435,
      "author_name": "muhammad4hmed",
      "author_url": "",
      "post_date": "07/17/2021 15:40:00",
      "content": "<p>Good work!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1396880,
      "author_name": "davidbroberts",
      "author_url": "",
      "post_date": "07/22/2021 14:58:56",
      "content": "<p>I made a dataset and a train notebook for base anatomy detection model.</p>\n<p>Dataset -&gt; <a href=\"https://www.kaggle.com/davidbroberts/chest-anatomy-yolov5-train-set\" target=\"_blank\">https://www.kaggle.com/davidbroberts/chest-anatomy-yolov5-train-set</a><br>\nTrain notebook -&gt; <a href=\"https://www.kaggle.com/davidbroberts/chest-anatomy-detection-train\" target=\"_blank\">https://www.kaggle.com/davidbroberts/chest-anatomy-detection-train</a></p>\n<p>Feel free to use it as you wish. The images are from the SIIM-COVID19 dataset, so the original license applies.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1385853": "I made a Torch/YOLOv5 model to detect lungs, clavicles, shoulders and the heart. The idea is to figure out where to crop in order to eliminate as much 'noise' as possible.\n\nI call it 'Smart Cropping'\n\nThis notebook demonstrates demonstrates predicting bounding boxes on anatomy, and then cropping the image to fit.\n\nhttps://www.kaggle.com/davidbroberts/smart-cropping-cxr-with-yolov5",
    "1386702": "I updated the anatomy detection model in this notebook to include the trachea, carina and gastric bubble.",
    "1391435": "Good work!",
    "1396880": "I made a dataset and a train notebook for base anatomy detection model.\n\nDataset -> https://www.kaggle.com/davidbroberts/chest-anatomy-yolov5-train-set\nTrain notebook -> https://www.kaggle.com/davidbroberts/chest-anatomy-detection-train\n\nFeel free to use it as you wish. The images are from the SIIM-COVID19 dataset, so the original license applies."
  },
  "source": "meta"
}