{
  "id": 159444,
  "title": "UNET Vs Other Models ?",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/159444",
  "author_name": "Muhammad Mohsin",
  "post_date": "2020-06-17T13:15:14.145000",
  "votes": 0,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello everyone  ,\nProstate cancer is the second most deadliest form of  cancer found in men. Due to large scale inter observer variation there is dire need of automation for correctly grading the prostate cancer .For this purpose , I have used UNET architecture for training and pixel level prediction and achieve Cohen's Kappa 0.77.  The dataset which I used is Tissue Microarrays (TMA) based RGB histological images. Now I am looking for another models which are also very useful in this problems , which my fellows tried here. Please share the model details with some useful steps so we can discuss and improve the problematic results in teams </p>",
  "messages": [
    {
      "id": 907954,
      "postDate": "2020-06-30T09:29:42.217Z",
      "content": "<p>Hi, have you tried to only train the model in radboud dataset? It seems only radboud dataset have full pixel-level mask</p>",
      "rawMarkdown": "Hi, have you tried to only train the model in radboud dataset? It seems only radboud dataset have full pixel-level mask"
    },
    {
      "id": 890355,
      "postDate": "2020-06-17T13:15:14.147Z",
      "content": "<p>Hello everyone  ,\nProstate cancer is the second most deadliest form of  cancer found in men. Due to large scale inter observer variation there is dire need of automation for correctly grading the prostate cancer .For this purpose , I have used UNET architecture for training and pixel level prediction and achieve Cohen's Kappa 0.77.  The dataset which I used is Tissue Microarrays (TMA) based RGB histological images. Now I am looking for another models which are also very useful in this problems , which my fellows tried here. Please share the model details with some useful steps so we can discuss and improve the problematic results in teams </p>",
      "rawMarkdown": "Hello everyone  ,\nProstate cancer is the second most deadliest form of  cancer found in men. Due to large scale inter observer variation there is dire need of automation for correctly grading the prostate cancer .For this purpose , I have used UNET architecture for training and pixel level prediction and achieve Cohen's Kappa 0.77.  The dataset which I used is Tissue Microarrays (TMA) based RGB histological images. Now I am looking for another models which are also very useful in this problems , which my fellows tried here. Please share the model details with some useful steps so we can discuss and improve the problematic results in teams "
    }
  ],
  "comments": [
    {
      "id": 907954,
      "author_name": "Shuolin Liu",
      "author_url": "",
      "post_date": "2020-06-30T09:29:42.217000",
      "content": "<p>Hi, have you tried to only train the model in radboud dataset? It seems only radboud dataset have full pixel-level mask</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "907954": "Hi, have you tried to only train the model in radboud dataset? It seems only radboud dataset have full pixel-level mask",
    "890355": "Hello everyone  ,\nProstate cancer is the second most deadliest form of  cancer found in men. Due to large scale inter observer variation there is dire need of automation for correctly grading the prostate cancer .For this purpose , I have used UNET architecture for training and pixel level prediction and achieve Cohen's Kappa 0.77.  The dataset which I used is Tissue Microarrays (TMA) based RGB histological images. Now I am looking for another models which are also very useful in this problems , which my fellows tried here. Please share the model details with some useful steps so we can discuss and improve the problematic results in teams "
  }
}