{
  "id": 379139,
  "title": "A question about an approach that can be used.",
  "url": "/competitions/nfl-player-contact-detection/discussion/379139",
  "author_name": "",
  "post_date": "2023-01-18T09:28:26.384576100Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Do You think it would be sufficient to use the baseline helmet bounding boxes only to train a model to detect contact with ground, without using the video data? <br>\nAnd maybe using algorithms such as yolo to do instance segmentation on the video data and use it to detect contact between different instances (players)?</p>",
  "messages": [
    {
      "id": "2105125",
      "postDate": "01/18/2023 09:28:26",
      "content": "<p>Do You think it would be sufficient to use the baseline helmet bounding boxes only to train a model to detect contact with ground, without using the video data? <br>\nAnd maybe using algorithms such as yolo to do instance segmentation on the video data and use it to detect contact between different instances (players)?</p>",
      "rawMarkdown": "Do You think it would be sufficient to use the baseline helmet bounding boxes only to train a model to detect contact with ground, without using the video data? \nAnd maybe using algorithms such as yolo to do instance segmentation on the video data and use it to detect contact between different instances (players)?",
      "votes": null
    },
    {
      "id": "2106282",
      "postDate": "01/19/2023 04:18:19",
      "content": "<p>It is definitely worth considering but at the end of the day there is only one way to find out..</p>\n<p>The Devastator.</p>",
      "rawMarkdown": "It is definitely worth considering but at the end of the day there is only one way to find out..\n\nThe Devastator.",
      "votes": null
    },
    {
      "id": "2106469",
      "postDate": "01/19/2023 07:04:13",
      "content": "<p>From what I'm gathering over the discussion topics, I think people are using much simpler ideas, since their submission time is low. But I could be wrong.</p>",
      "rawMarkdown": "From what I'm gathering over the discussion topics, I think people are using much simpler ideas, since their submission time is low. But I could be wrong.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2106282,
      "author_name": "thedevastator",
      "author_url": "",
      "post_date": "01/19/2023 04:18:19",
      "content": "<p>It is definitely worth considering but at the end of the day there is only one way to find out..</p>\n<p>The Devastator.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2106469,
      "author_name": "samir95",
      "author_url": "",
      "post_date": "01/19/2023 07:04:13",
      "content": "<p>From what I'm gathering over the discussion topics, I think people are using much simpler ideas, since their submission time is low. But I could be wrong.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2105125": "Do You think it would be sufficient to use the baseline helmet bounding boxes only to train a model to detect contact with ground, without using the video data? \nAnd maybe using algorithms such as yolo to do instance segmentation on the video data and use it to detect contact between different instances (players)?",
    "2106282": "It is definitely worth considering but at the end of the day there is only one way to find out..\n\nThe Devastator.",
    "2106469": "From what I'm gathering over the discussion topics, I think people are using much simpler ideas, since their submission time is low. But I could be wrong."
  },
  "source": "meta"
}