{
  "id": 535066,
  "title": "I need clarification on submission.csv",
  "url": "/competitions/rsna-2024-lumbar-spine-degenerative-classification/discussion/535066",
  "author_name": "",
  "post_date": "2024-09-20T06:25:02.787857600Z",
  "votes": -1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hello, I am struggling with how the test data should be sent. The folder inside \"Test\" contains Study_IDs, which are used as the file names. However, the study_ID alone does not indicate which image is being tested. Could we add the siries_ID and instance number to the file name as well? This way, it becomes clear which image is being tested.</p>",
  "messages": [
    {
      "id": "2993691",
      "postDate": "09/20/2024 06:25:02",
      "content": "<p>Hello, I am struggling with how the test data should be sent. The folder inside \"Test\" contains Study_IDs, which are used as the file names. However, the study_ID alone does not indicate which image is being tested. Could we add the siries_ID and instance number to the file name as well? This way, it becomes clear which image is being tested.</p>",
      "rawMarkdown": "Hello, I am struggling with how the test data should be sent. The folder inside \"Test\" contains Study_IDs, which are used as the file names. However, the study_ID alone does not indicate which image is being tested. Could we add the siries_ID and instance number to the file name as well? This way, it becomes clear which image is being tested.",
      "votes": null
    },
    {
      "id": "2993746",
      "postDate": "09/20/2024 07:48:47",
      "content": "<p>row_id's are {study_id}<em>{patology}</em>{level}</p>\n<p>For each study_id you have to submit the corresponding 25 combination of 5 patology x 5 level row_id, and for each row_id the corresponding 3 probabilies of being that row 'Normal/Mild', 'Moderate' or 'Severe'.</p>\n<p>At \"test_series_descriptions.csv\" you have the information of what series_id corresponds to each study_id. That's what series_id's you will find inside the study_id's folders and what type of MRI are.</p>\n<p>To finish, we need to submit everything for each case (study_id) independently of the information available. Is not a prediction per MRI (series_id).</p>",
      "rawMarkdown": "row_id's are {study_id}_{patology}_{level}\n\nFor each study_id you have to submit the corresponding 25 combination of 5 patology x 5 level row_id, and for each row_id the corresponding 3 probabilies of being that row 'Normal/Mild', 'Moderate' or 'Severe'.\n\nAt \"test_series_descriptions.csv\" you have the information of what series_id corresponds to each study_id. That's what series_id's you will find inside the study_id's folders and what type of MRI are.\n\nTo finish, we need to submit everything for each case (study_id) independently of the information available. Is not a prediction per MRI (series_id).",
      "votes": null
    },
    {
      "id": "2993751",
      "postDate": "09/20/2024 07:55:17",
      "content": "<p>I know, but when I want to test. I am giving one image to the model. and it should give me condition and label of it. which will be 25 ids for each image. what I am saying is that in the submission.csv we are just pointing out to the study_id of the image not the siries_id and instance number. so how our image is identified ? for example which of these images is tested? </p>",
      "rawMarkdown": "I know, but when I want to test. I am giving one image to the model. and it should give me condition and label of it. which will be 25 ids for each image. what I am saying is that in the submission.csv we are just pointing out to the study_id of the image not the siries_id and instance number. so how our image is identified ? for example which of these images is tested?",
      "votes": null
    },
    {
      "id": "2993775",
      "postDate": "09/20/2024 08:19:09",
      "content": "<p>Because is pointing only the study_id, and has his all corresponding series_id available. Again, \"Is not a prediction per MRI image (series_id).\" All study_id folder is beign tested, what images to use is our job.</p>\n<p>EDIT: <a href=\"https://www.kaggle.com/code/abhinavsuri/anatomy-image-visualization-overview-rsna-raids\" target=\"_blank\">https://www.kaggle.com/code/abhinavsuri/anatomy-image-visualization-overview-rsna-raids</a></p>",
      "rawMarkdown": "Because is pointing only the study_id, and has his all corresponding series_id available. Again, \"Is not a prediction per MRI image (series_id).\" All study_id folder is beign tested, what images to use is our job.\n\nEDIT: https://www.kaggle.com/code/abhinavsuri/anatomy-image-visualization-overview-rsna-raids",
      "votes": null
    },
    {
      "id": "3008641",
      "postDate": "10/06/2024 22:23:06",
      "content": "<p>but how do you associate each image to their corresponding pathology and level to further predict each probability for 'Normal/Mild', 'Moderate' or 'Severe` ?</p>",
      "rawMarkdown": "but how do you associate each image to their corresponding pathology and level to further predict each probability for 'Normal/Mild', 'Moderate' or 'Severe` ?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2993746,
      "author_name": "sacuscreed",
      "author_url": "",
      "post_date": "09/20/2024 07:48:47",
      "content": "<p>row_id's are {study_id}<em>{patology}</em>{level}</p>\n<p>For each study_id you have to submit the corresponding 25 combination of 5 patology x 5 level row_id, and for each row_id the corresponding 3 probabilies of being that row 'Normal/Mild', 'Moderate' or 'Severe'.</p>\n<p>At \"test_series_descriptions.csv\" you have the information of what series_id corresponds to each study_id. That's what series_id's you will find inside the study_id's folders and what type of MRI are.</p>\n<p>To finish, we need to submit everything for each case (study_id) independently of the information available. Is not a prediction per MRI (series_id).</p>",
      "votes": null,
      "replies": [
        {
          "id": 2993751,
          "author_name": "saminifard",
          "author_url": "",
          "post_date": "09/20/2024 07:55:17",
          "content": "<p>I know, but when I want to test. I am giving one image to the model. and it should give me condition and label of it. which will be 25 ids for each image. what I am saying is that in the submission.csv we are just pointing out to the study_id of the image not the siries_id and instance number. so how our image is identified ? for example which of these images is tested? </p>",
          "votes": null,
          "replies": [
            {
              "id": 2993775,
              "author_name": "sacuscreed",
              "author_url": "",
              "post_date": "09/20/2024 08:19:09",
              "content": "<p>Because is pointing only the study_id, and has his all corresponding series_id available. Again, \"Is not a prediction per MRI image (series_id).\" All study_id folder is beign tested, what images to use is our job.</p>\n<p>EDIT: <a href=\"https://www.kaggle.com/code/abhinavsuri/anatomy-image-visualization-overview-rsna-raids\" target=\"_blank\">https://www.kaggle.com/code/abhinavsuri/anatomy-image-visualization-overview-rsna-raids</a></p>",
              "votes": null,
              "replies": [
                {
                  "id": 3008641,
                  "author_name": "evandrocardozo",
                  "author_url": "",
                  "post_date": "10/06/2024 22:23:06",
                  "content": "<p>but how do you associate each image to their corresponding pathology and level to further predict each probability for 'Normal/Mild', 'Moderate' or 'Severe` ?</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2993691": "Hello, I am struggling with how the test data should be sent. The folder inside \"Test\" contains Study_IDs, which are used as the file names. However, the study_ID alone does not indicate which image is being tested. Could we add the siries_ID and instance number to the file name as well? This way, it becomes clear which image is being tested.",
    "2993746": "row_id's are {study_id}_{patology}_{level}\n\nFor each study_id you have to submit the corresponding 25 combination of 5 patology x 5 level row_id, and for each row_id the corresponding 3 probabilies of being that row 'Normal/Mild', 'Moderate' or 'Severe'.\n\nAt \"test_series_descriptions.csv\" you have the information of what series_id corresponds to each study_id. That's what series_id's you will find inside the study_id's folders and what type of MRI are.\n\nTo finish, we need to submit everything for each case (study_id) independently of the information available. Is not a prediction per MRI (series_id).",
    "2993751": "I know, but when I want to test. I am giving one image to the model. and it should give me condition and label of it. which will be 25 ids for each image. what I am saying is that in the submission.csv we are just pointing out to the study_id of the image not the siries_id and instance number. so how our image is identified ? for example which of these images is tested?",
    "2993775": "Because is pointing only the study_id, and has his all corresponding series_id available. Again, \"Is not a prediction per MRI image (series_id).\" All study_id folder is beign tested, what images to use is our job.\n\nEDIT: https://www.kaggle.com/code/abhinavsuri/anatomy-image-visualization-overview-rsna-raids",
    "3008641": "but how do you associate each image to their corresponding pathology and level to further predict each probability for 'Normal/Mild', 'Moderate' or 'Severe` ?"
  },
  "source": "meta"
}