{
  "id": 240280,
  "title": "Use of Segmentation First then Classification!",
  "url": "/competitions/siim-covid19-detection/discussion/240280",
  "author_name": "",
  "post_date": "2021-05-19T06:25:55.787567700Z",
  "votes": 24,
  "comment_count": 2,
  "views": 0,
  "content": "<p>My approach to this deep learning problem would be to first apply some Segmentation Models to extract Lung masks and then perform Classification over the same.</p>\n<p>This approach would help in solving neural network overfitting as mentioned in these papers: </p>\n<ul>\n<li>Unveiling COVID-19 from Chest X-ray with deep learning: a hurdles race with small data Enzo Tartaglione, Carlo Alberto Barbano, Claudio Berzovini, Marco Calandri, Marco Grangetto<br>\n<a href=\"https://arxiv.org/abs/2004.05405\" target=\"_blank\">https://arxiv.org/abs/2004.05405</a></li>\n<li>A Critic Evaluation of Methods for COVID-19 Automatic Detection from X-Ray Images Gianluca Maguolo, Loris Nanni, <a href=\"https://arxiv.org/abs/2004.12823\" target=\"_blank\">https://arxiv.org/abs/2004.12823</a></li>\n</ul>\n<p>What are your thoughts on the same?</p>",
  "messages": [
    {
      "id": "1314391",
      "postDate": "05/19/2021 06:25:55",
      "content": "<p>My approach to this deep learning problem would be to first apply some Segmentation Models to extract Lung masks and then perform Classification over the same.</p>\n<p>This approach would help in solving neural network overfitting as mentioned in these papers: </p>\n<ul>\n<li>Unveiling COVID-19 from Chest X-ray with deep learning: a hurdles race with small data Enzo Tartaglione, Carlo Alberto Barbano, Claudio Berzovini, Marco Calandri, Marco Grangetto<br>\n<a href=\"https://arxiv.org/abs/2004.05405\" target=\"_blank\">https://arxiv.org/abs/2004.05405</a></li>\n<li>A Critic Evaluation of Methods for COVID-19 Automatic Detection from X-Ray Images Gianluca Maguolo, Loris Nanni, <a href=\"https://arxiv.org/abs/2004.12823\" target=\"_blank\">https://arxiv.org/abs/2004.12823</a></li>\n</ul>\n<p>What are your thoughts on the same?</p>",
      "rawMarkdown": "My approach to this deep learning problem would be to first apply some Segmentation Models to extract Lung masks and then perform Classification over the same.\n\nThis approach would help in solving neural network overfitting as mentioned in these papers: \n- Unveiling COVID-19 from Chest X-ray with deep learning: a hurdles race with small data Enzo Tartaglione, Carlo Alberto Barbano, Claudio Berzovini, Marco Calandri, Marco Grangetto\nhttps://arxiv.org/abs/2004.05405\n- A Critic Evaluation of Methods for COVID-19 Automatic Detection from X-Ray Images Gianluca Maguolo, Loris Nanni, https://arxiv.org/abs/2004.12823\n\nWhat are your thoughts on the same?",
      "votes": null
    },
    {
      "id": "1315487",
      "postDate": "05/19/2021 20:43:35",
      "content": "<p>Resources :</p>\n<p>Pretrained UNET Lung Segmentation Dataset  Notebook: <a href=\"https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-dataset\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-dataset</a></p>\n<p>UNET Lung Segmentation Dataset SIIM COVID : <a href=\"https://www.kaggle.com/farhanhaikhan/unet-lung-segmentation-dataset-siim-covid/code\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/unet-lung-segmentation-dataset-siim-covid/code</a></p>\n<p>MaskEDAandFiltration :<br>\n<a href=\"https://www.kaggle.com/farhanhaikhan/maskedaandfiltration/comments\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/maskedaandfiltration/comments</a></p>\n<p>Pretrained UNET Lung Segmentation MetaData:<br>\n<a href=\"https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-metadata\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-metadata</a></p>",
      "rawMarkdown": "Resources :\n\nPretrained UNET Lung Segmentation Dataset  Notebook: https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-dataset\n\nUNET Lung Segmentation Dataset SIIM COVID : https://www.kaggle.com/farhanhaikhan/unet-lung-segmentation-dataset-siim-covid/code\n\nMaskEDAandFiltration :\nhttps://www.kaggle.com/farhanhaikhan/maskedaandfiltration/comments\n\nPretrained UNET Lung Segmentation MetaData:\nhttps://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-metadata",
      "votes": null
    },
    {
      "id": "1394880",
      "postDate": "07/20/2021 17:36:33",
      "content": "<p>Hello. I want to try segmentation for training a classifier. Are you training the model using these masks?</p>",
      "rawMarkdown": "Hello. I want to try segmentation for training a classifier. Are you training the model using these masks?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1315487,
      "author_name": "farhanhaikhan",
      "author_url": "",
      "post_date": "05/19/2021 20:43:35",
      "content": "<p>Resources :</p>\n<p>Pretrained UNET Lung Segmentation Dataset  Notebook: <a href=\"https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-dataset\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-dataset</a></p>\n<p>UNET Lung Segmentation Dataset SIIM COVID : <a href=\"https://www.kaggle.com/farhanhaikhan/unet-lung-segmentation-dataset-siim-covid/code\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/unet-lung-segmentation-dataset-siim-covid/code</a></p>\n<p>MaskEDAandFiltration :<br>\n<a href=\"https://www.kaggle.com/farhanhaikhan/maskedaandfiltration/comments\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/maskedaandfiltration/comments</a></p>\n<p>Pretrained UNET Lung Segmentation MetaData:<br>\n<a href=\"https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-metadata\" target=\"_blank\">https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-metadata</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1394880,
      "author_name": "zaakciiru",
      "author_url": "",
      "post_date": "07/20/2021 17:36:33",
      "content": "<p>Hello. I want to try segmentation for training a classifier. Are you training the model using these masks?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1314391": "My approach to this deep learning problem would be to first apply some Segmentation Models to extract Lung masks and then perform Classification over the same.\n\nThis approach would help in solving neural network overfitting as mentioned in these papers: \n- Unveiling COVID-19 from Chest X-ray with deep learning: a hurdles race with small data Enzo Tartaglione, Carlo Alberto Barbano, Claudio Berzovini, Marco Calandri, Marco Grangetto\nhttps://arxiv.org/abs/2004.05405\n- A Critic Evaluation of Methods for COVID-19 Automatic Detection from X-Ray Images Gianluca Maguolo, Loris Nanni, https://arxiv.org/abs/2004.12823\n\nWhat are your thoughts on the same?",
    "1315487": "Resources :\n\nPretrained UNET Lung Segmentation Dataset  Notebook: https://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-dataset\n\nUNET Lung Segmentation Dataset SIIM COVID : https://www.kaggle.com/farhanhaikhan/unet-lung-segmentation-dataset-siim-covid/code\n\nMaskEDAandFiltration :\nhttps://www.kaggle.com/farhanhaikhan/maskedaandfiltration/comments\n\nPretrained UNET Lung Segmentation MetaData:\nhttps://www.kaggle.com/farhanhaikhan/pretrained-unet-lung-segmentation-metadata",
    "1394880": "Hello. I want to try segmentation for training a classifier. Are you training the model using these masks?"
  },
  "source": "meta"
}