![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F609324%2F9478c8ddc0efe1aa59e58c6b68909dff%2F5.PNG?generation=1569261394240994&alt=media\")
Hello,
\n\nDoes the labeling is based on a particular image, or is it the same for all the slice of the same patient ?
\n\nTank you !
", "messages": [ { "id": 631823, "postDate": "2019-09-22T18:47:06.760Z", "content": "Hello,
\n\nDoes the labeling is based on a particular image, or is it the same for all the slice of the same patient ?
\n\nTank you !
", "rawMarkdown": "Hello,\n\nDoes the labeling is based on a particular image, or is it the same for all the slice of the same patient ?\n\nTank you !", "votes": 6 }, { "id": 632451, "postDate": "2019-09-23T15:48:43.077Z", "content": "Hi all - thanks for your questions! The labeling is based on the particular image, as some others have posted here.
", "rawMarkdown": "Hi all - thanks for your questions! The labeling is based on the particular image, as some others have posted here.", "votes": 3 }, { "id": 631872, "postDate": "2019-09-22T21:03:23.170Z", "content": "I checked that, and it differed within the patients. So to answer your question: the labeling seems to be based on a particular image. The CT slices of a patient could have different labels.
", "rawMarkdown": "I checked that, and it differed within the patients. So to answer your question: the labeling seems to be based on a particular image. The CT slices of a patient could have different labels.", "votes": 3 }, { "id": 631837, "postDate": "2019-09-22T19:33:43.310Z", "content": "I agree that this needs clarification. Also, does each study have the same number of images?
", "rawMarkdown": "I agree that this needs clarification. Also, does each study have the same number of images? ", "votes": 1, "replies": [ { "id": 632001, "postDate": "2019-09-23T05:02:20.453Z", "content": "Nope, some have more than 100 images and some have only ~20 images.
", "rawMarkdown": "Nope, some have more than 100 images and some have only ~20 images." }, { "id": 632450, "postDate": "2019-09-23T15:48:21.610Z", "content": "Hi @mrhuangchao - thanks! Which study has more than 100 images?
", "rawMarkdown": "Hi @mrhuangchao - thanks! Which study has more than 100 images?", "votes": 1 }, { "id": 632488, "postDate": "2019-09-23T16:56:33.277Z", "content": "@philculliton Of the 17079 patients in the training set I've counted 365 and there are some with more than 200. This is the case for patients with different studies.\nSome examples:\nID_00a3b735, ID_0369d8b8, ID_03962d75
", "rawMarkdown": "@philculliton Of the 17079 patients in the training set I've counted 365 and there are some with more than 200. This is the case for patients with different studies.\nSome examples:\nID_00a3b735, ID_0369d8b8, ID_03962d75", "votes": 1 }, { "id": 632502, "postDate": "2019-09-23T17:18:31.890Z", "content": "@amyaramine - Thanks! Some patients do have multiple studies (as in these cases). No studies, on their own, should contain more than 60 images, though. The Study Instance UID defines what's contained within a particular study.
", "rawMarkdown": "@amyaramine - Thanks! Some patients do have multiple studies (as in these cases). No studies, on their own, should contain more than 60 images, though. The Study Instance UID defines what's contained within a particular study.", "votes": 2 }, { "id": 632524, "postDate": "2019-09-23T17:49:20.813Z", "content": "Agree ! I was able to find them using Study Instance UID ! However, I checked visually some of them and could not find any major changes -except- the image orientation and differences in the Image Patient Position !
", "rawMarkdown": "Agree ! I was able to find them using Study Instance UID ! However, I checked visually some of them and could not find any major changes -except- the image orientation and differences in the Image Patient Position !", "votes": 1 }, { "id": 632534, "postDate": "2019-09-23T17:56:19.327Z", "content": "Thank you Phil! This is an exceptional data set!
\n\nDoes anyone know if there are scans of totally healthy subjects? I'll be looking into this tonight and will report back.
", "rawMarkdown": "Thank you Phil! This is an exceptional data set!\n\nDoes anyone know if there are scans of totally healthy subjects? I'll be looking into this tonight and will report back." }, { "id": 632535, "postDate": "2019-09-23T17:57:15.083Z", "content": "For example:\n
Also, sometimes the difference is the number of slices in each study !
", "rawMarkdown": "For example:\n\n\nAlso, sometimes the difference is the number of slices in each study !", "votes": 2 }, { "id": 632538, "postDate": "2019-09-23T18:03:10.773Z", "content": "@thavlik There are healthy subjects ! check ID_00a3b735 for an example.\nPs: by the way this patient has 4 Study Instance UID !
", "rawMarkdown": "@thavlik There are healthy subjects ! check ID_00a3b735 for an example.\nPs: by the way this patient has 4 Study Instance UID !", "votes": 1 } ] }, { "id": 632309, "postDate": "2019-09-23T13:00:51.847Z", "content": "The labeling is done slice by slice ! For example the same patient may have 37 slices: some are labeled as zero (nothing to report) and some are labeled with the type of ICH (if it's positive). So, only the slices where we can see the hemorrhage are labeled as positive.
", "rawMarkdown": "The labeling is done slice by slice ! For example the same patient may have 37 slices: some are labeled as zero (nothing to report) and some are labeled with the type of ICH (if it's positive). So, only the slices where we can see the hemorrhage are labeled as positive.", "votes": 2 }, { "id": 835271, "postDate": "2020-05-06T06:44:39.427Z", "content": "could pls tell standford data set how to label it ,i just downloaded from the website not in kaggle
", "rawMarkdown": "could pls tell standford data set how to label it ,i just downloaded from the website not in kaggle\n" } ], "comments": [ { "id": 632451, "author_name": "Phil Culliton", "author_url": "", "post_date": "2019-09-23T15:48:43.077000", "content": "Hi all - thanks for your questions! The labeling is based on the particular image, as some others have posted here.
", "votes": 3, "replies": [] }, { "id": 631872, "author_name": "Alex", "author_url": "", "post_date": "2019-09-22T21:03:23.170000", "content": "I checked that, and it differed within the patients. So to answer your question: the labeling seems to be based on a particular image. The CT slices of a patient could have different labels.
", "votes": 3, "replies": [] }, { "id": 631837, "author_name": "Tom H.", "author_url": "", "post_date": "2019-09-22T19:33:43.310000", "content": "I agree that this needs clarification. Also, does each study have the same number of images?
", "votes": 1, "replies": [ { "id": 632001, "author_name": "Chao Huang", "author_url": "", "post_date": "2019-09-23T05:02:20.453000", "content": "Nope, some have more than 100 images and some have only ~20 images.
", "votes": 0, "replies": [] }, { "id": 632450, "author_name": "Phil Culliton", "author_url": "", "post_date": "2019-09-23T15:48:21.610000", "content": "Hi @mrhuangchao - thanks! Which study has more than 100 images?
", "votes": 1, "replies": [] }, { "id": 632488, "author_name": "Al-Khwârizmî", "author_url": "", "post_date": "2019-09-23T16:56:33.277000", "content": "@philculliton Of the 17079 patients in the training set I've counted 365 and there are some with more than 200. This is the case for patients with different studies.\nSome examples:\nID_00a3b735, ID_0369d8b8, ID_03962d75
", "votes": 1, "replies": [] }, { "id": 632502, "author_name": "Phil Culliton", "author_url": "", "post_date": "2019-09-23T17:18:31.890000", "content": "@amyaramine - Thanks! Some patients do have multiple studies (as in these cases). No studies, on their own, should contain more than 60 images, though. The Study Instance UID defines what's contained within a particular study.
", "votes": 2, "replies": [] }, { "id": 632524, "author_name": "Al-Khwârizmî", "author_url": "", "post_date": "2019-09-23T17:49:20.813000", "content": "Agree ! I was able to find them using Study Instance UID ! However, I checked visually some of them and could not find any major changes -except- the image orientation and differences in the Image Patient Position !
", "votes": 1, "replies": [] }, { "id": 632534, "author_name": "Tom H.", "author_url": "", "post_date": "2019-09-23T17:56:19.327000", "content": "Thank you Phil! This is an exceptional data set!
\n\nDoes anyone know if there are scans of totally healthy subjects? I'll be looking into this tonight and will report back.
", "votes": 0, "replies": [] }, { "id": 632535, "author_name": "Al-Khwârizmî", "author_url": "", "post_date": "2019-09-23T17:57:15.083000", "content": "For example:\n
Also, sometimes the difference is the number of slices in each study !
", "votes": 2, "replies": [] }, { "id": 632538, "author_name": "Al-Khwârizmî", "author_url": "", "post_date": "2019-09-23T18:03:10.773000", "content": "@thavlik There are healthy subjects ! check ID_00a3b735 for an example.\nPs: by the way this patient has 4 Study Instance UID !
", "votes": 1, "replies": [] } ] }, { "id": 632309, "author_name": "Al-Khwârizmî", "author_url": "", "post_date": "2019-09-23T13:00:51.847000", "content": "The labeling is done slice by slice ! For example the same patient may have 37 slices: some are labeled as zero (nothing to report) and some are labeled with the type of ICH (if it's positive). So, only the slices where we can see the hemorrhage are labeled as positive.
", "votes": 2, "replies": [] }, { "id": 835271, "author_name": "Selvaraj B", "author_url": "", "post_date": "2020-05-06T06:44:39.427000", "content": "could pls tell standford data set how to label it ,i just downloaded from the website not in kaggle
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "631823": "Hello,\n\nDoes the labeling is based on a particular image, or is it the same for all the slice of the same patient ?\n\nTank you !", "632451": "Hi all - thanks for your questions! The labeling is based on the particular image, as some others have posted here.", "631872": "I checked that, and it differed within the patients. So to answer your question: the labeling seems to be based on a particular image. The CT slices of a patient could have different labels.", "631837": "I agree that this needs clarification. Also, does each study have the same number of images? ", "632309": "The labeling is done slice by slice ! For example the same patient may have 37 slices: some are labeled as zero (nothing to report) and some are labeled with the type of ICH (if it's positive). So, only the slices where we can see the hemorrhage are labeled as positive.", "835271": "could pls tell standford data set how to label it ,i just downloaded from the website not in kaggle\n" } }