{
  "id": 663516,
  "title": "How are you extracting the waveforms?",
  "url": "/competitions/physionet-ecg-image-digitization/discussion/663516",
  "author_name": "",
  "post_date": "2025-12-18T12:51:10.057138600Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I am currently using my own manually annotated images as training data and extracting the waveforms with a U-Net–based model.</p>\n<p>If you don’t mind sharing, I’d love to hear how others are extracting waveforms and what models or approaches you are using.</p>",
  "messages": [
    {
      "id": "3378537",
      "postDate": "12/18/2025 12:51:10",
      "content": "<p>I am currently using my own manually annotated images as training data and extracting the waveforms with a U-Net–based model.</p>\n<p>If you don’t mind sharing, I’d love to hear how others are extracting waveforms and what models or approaches you are using.</p>",
      "rawMarkdown": "I am currently using my own manually annotated images as training data and extracting the waveforms with a U-Net–based model.\n\nIf you don’t mind sharing, I’d love to hear how others are extracting waveforms and what models or approaches you are using.",
      "votes": null
    },
    {
      "id": "3379703",
      "postDate": "12/20/2025 13:07:30",
      "content": "<p>I think the most natural approach is to start with a U-net based model, as it was the winner of the past competition:\n<a href=\"https://github.com/felixkrones/ECG-Digitiser/tree/main\" target=\"_blank\">https://github.com/felixkrones/ECG-Digitiser/tree/main</a></p>\n<p>While I will try something similar, I cannot deny that I also want to find another approach.</p>",
      "rawMarkdown": "I think the most natural approach is to start with a U-net based model, as it was the winner of the past competition:\nhttps://github.com/felixkrones/ECG-Digitiser/tree/main\n\nWhile I will try something similar, I cannot deny that I also want to find another approach.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3379703,
      "author_name": "juancneira",
      "author_url": "",
      "post_date": "12/20/2025 13:07:30",
      "content": "<p>I think the most natural approach is to start with a U-net based model, as it was the winner of the past competition:\n<a href=\"https://github.com/felixkrones/ECG-Digitiser/tree/main\" target=\"_blank\">https://github.com/felixkrones/ECG-Digitiser/tree/main</a></p>\n<p>While I will try something similar, I cannot deny that I also want to find another approach.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3378537": "I am currently using my own manually annotated images as training data and extracting the waveforms with a U-Net–based model.\n\nIf you don’t mind sharing, I’d love to hear how others are extracting waveforms and what models or approaches you are using.",
    "3379703": "I think the most natural approach is to start with a U-net based model, as it was the winner of the past competition:\nhttps://github.com/felixkrones/ECG-Digitiser/tree/main\n\nWhile I will try something similar, I cannot deny that I also want to find another approach."
  },
  "source": "meta"
}