![\"https://i.ibb.co/HnBdscc/Selection-622.png\"](\"https://i.ibb.co/HnBdscc/Selection-622.png\")
for more information, you should refer to the dataset and paper here:
\nhttps://bimcv.cipf.es/bimcv-projects/bimcv-covid19/
\n(e.g. accuracy of the labels, what labels are provided, etc)
here i showed an example image from the extenal bimcv-covid19 dataset.
\nthe disease label and location label are provided
you should refer to the file : /covid19_posi/derivatives/labels/labels_covid_posi.tsv
\n(feature,location) pair seems good for pretraining transformer
", "messages": [ { "id": 1400789, "postDate": "2021-07-26T15:21:13.150Z", "content": "
for more information, you should refer to the dataset and paper here:
\nhttps://bimcv.cipf.es/bimcv-projects/bimcv-covid19/
\n(e.g. accuracy of the labels, what labels are provided, etc)
here i showed an example image from the extenal bimcv-covid19 dataset.
\nthe disease label and location label are provided
you should refer to the file : /covid19_posi/derivatives/labels/labels_covid_posi.tsv
\n(feature,location) pair seems good for pretraining transformer
", "rawMarkdown": "\n\nfor more information, you should refer to the dataset and paper here:\nhttps://bimcv.cipf.es/bimcv-projects/bimcv-covid19/\n(e.g. accuracy of the labels, what labels are provided, etc)\n\nhere i showed an example image from the extenal bimcv-covid19 dataset.\nthe disease label and location label are provided\n\nyou should refer to the file : /covid19_posi/derivatives/labels/labels_covid_posi.tsv\n\n(feature,location) pair seems good for pretraining transformer", "votes": 21 }, { "id": 1401353, "postDate": "2021-07-27T08:52:36.080Z", "content": "512x512 resized external data images (for bimcv covid +ve cases only) and list of about 5300 matched images (kaggle and bimcv) can be found here:
\nhttps://drive.google.com/drive/folders/1cCovVVwKSIiz9v62sRrBzhAzwVYlB3r-?usp=sharing
\n![\"https://i.ibb.co/09N72Mt/Selection-631.png\"](\"https://i.ibb.co/09N72Mt/Selection-631.png\")
thanks for sharing @hengck23. I have a question about the label of each study, I see the raw label from BIMCV doesn't contain \"negative\", \"typical\", \"indeterminate\" or \"atypical\". How do you get the label like that?
", "rawMarkdown": "thanks for sharing @hengck23. I have a question about the label of each study, I see the raw label from BIMCV doesn't contain \"negative\", \"typical\", \"indeterminate\" or \"atypical\". How do you get the label like that?" } ] }, { "id": 1403430, "postDate": "2021-07-29T05:44:25.483Z", "content": "Has anyone succeeded in using external dataset?
", "rawMarkdown": "Has anyone succeeded in using external dataset?", "votes": 1, "replies": [ { "id": 1403988, "postDate": "2021-07-29T13:29:48.613Z", "content": "I think the first place may use bimcv😊 MIDRC-RICORD also improved my study level score a bit.
", "rawMarkdown": "I think the first place may use bimcv😊 MIDRC-RICORD also improved my study level score a bit. ", "votes": 2 } ] }, { "id": 1400932, "postDate": "2021-07-26T18:08:40.270Z", "content": "Good work! Can you please give your opinion to my question https://www.kaggle.com/c/siim-covid19-detection/discussion/255283 ary if it's obvious, as i am very new to this platform.
", "rawMarkdown": "Good work! Can you please give your opinion to my question [https://www.kaggle.com/c/siim-covid19-detection/discussion/255283](url) ary if it's obvious, as i am very new to this platform." }, { "id": 1422653, "postDate": "2021-08-03T05:03:41.327Z", "content": "I think we can get good results for study level in identifying negative and non negative. For this we can use semi supervised learning. But for identifying among atypical and indeterminate cases, they are ambiguous. Extra unlabelled data may or may not help
", "rawMarkdown": "I think we can get good results for study level in identifying negative and non negative. For this we can use semi supervised learning. But for identifying among atypical and indeterminate cases, they are ambiguous. Extra unlabelled data may or may not help" }, { "id": 1402794, "postDate": "2021-07-28T14:38:06.173Z", "content": "Great share. Extra data is useful for the competition.
", "rawMarkdown": "Great share. Extra data is useful for the competition." }, { "id": 1402688, "postDate": "2021-07-28T12:51:42.743Z", "content": "i may be wrong. but after studying the bimcv and kaggle dataset, I think:
\nThanks for sharing. This is helpful.
", "rawMarkdown": "Thanks for sharing. This is helpful.", "votes": 1 }, { "id": 1401318, "postDate": "2021-07-27T07:56:58.970Z", "content": "Thanks for sharing @hengck23
", "rawMarkdown": "Thanks for sharing @hengck23 ", "votes": 1 } ], "comments": [ { "id": 1401353, "author_name": "hengck23", "author_url": "", "post_date": "2021-07-27T08:52:36.080000", "content": "512x512 resized external data images (for bimcv covid +ve cases only) and list of about 5300 matched images (kaggle and bimcv) can be found here:
\nhttps://drive.google.com/drive/folders/1cCovVVwKSIiz9v62sRrBzhAzwVYlB3r-?usp=sharing
\n
thanks for sharing @hengck23. I have a question about the label of each study, I see the raw label from BIMCV doesn't contain \"negative\", \"typical\", \"indeterminate\" or \"atypical\". How do you get the label like that?
", "votes": 0, "replies": [] } ] }, { "id": 1403430, "author_name": "cool_rabbit", "author_url": "", "post_date": "2021-07-29T05:44:25.483000", "content": "Has anyone succeeded in using external dataset?
", "votes": 1, "replies": [ { "id": 1403988, "author_name": "erkut", "author_url": "", "post_date": "2021-07-29T13:29:48.613000", "content": "I think the first place may use bimcv😊 MIDRC-RICORD also improved my study level score a bit.
", "votes": 2, "replies": [] } ] }, { "id": 1400932, "author_name": "Samrendra Roy", "author_url": "", "post_date": "2021-07-26T18:08:40.270000", "content": "Good work! Can you please give your opinion to my question https://www.kaggle.com/c/siim-covid19-detection/discussion/255283 ary if it's obvious, as i am very new to this platform.
", "votes": 0, "replies": [] }, { "id": 1422653, "author_name": "hengck23", "author_url": "", "post_date": "2021-08-03T05:03:41.327000", "content": "I think we can get good results for study level in identifying negative and non negative. For this we can use semi supervised learning. But for identifying among atypical and indeterminate cases, they are ambiguous. Extra unlabelled data may or may not help
", "votes": 0, "replies": [] }, { "id": 1402794, "author_name": "YOLO.H", "author_url": "", "post_date": "2021-07-28T14:38:06.173000", "content": "Great share. Extra data is useful for the competition.
", "votes": 0, "replies": [] }, { "id": 1402688, "author_name": "hengck23", "author_url": "", "post_date": "2021-07-28T12:51:42.743000", "content": "i may be wrong. but after studying the bimcv and kaggle dataset, I think:
\nThanks for sharing. This is helpful.
", "votes": 1, "replies": [] }, { "id": 1401318, "author_name": "Dr. Amritpal Singh", "author_url": "", "post_date": "2021-07-27T07:56:58.970000", "content": "Thanks for sharing @hengck23
", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "1400789": "\n\nfor more information, you should refer to the dataset and paper here:\nhttps://bimcv.cipf.es/bimcv-projects/bimcv-covid19/\n(e.g. accuracy of the labels, what labels are provided, etc)\n\nhere i showed an example image from the extenal bimcv-covid19 dataset.\nthe disease label and location label are provided\n\nyou should refer to the file : /covid19_posi/derivatives/labels/labels_covid_posi.tsv\n\n(feature,location) pair seems good for pretraining transformer", "1401353": "512x512 resized external data images (for bimcv covid +ve cases only) and list of about 5300 matched images (kaggle and bimcv) can be found here:\n\nhttps://drive.google.com/drive/folders/1cCovVVwKSIiz9v62sRrBzhAzwVYlB3r-?usp=sharing\n\n", "1403430": "Has anyone succeeded in using external dataset?", "1400932": "Good work! Can you please give your opinion to my question [https://www.kaggle.com/c/siim-covid19-detection/discussion/255283](url) ary if it's obvious, as i am very new to this platform.", "1422653": "I think we can get good results for study level in identifying negative and non negative. For this we can use semi supervised learning. But for identifying among atypical and indeterminate cases, they are ambiguous. Extra unlabelled data may or may not help", "1402794": "Great share. Extra data is useful for the competition.", "1402688": "i may be wrong. but after studying the bimcv and kaggle dataset, I think:\n\n- for kaggle train dataset, one study has one grading label (negative, typical, indeterminate, atypical)\n- some study has a few sessions, i.e. a few images.\n- from bimcv, these sessions are taken at different times. and they have different labels.\n- hence the label of the study for the kaggle train dataset may only refer to one (or some, but not necessary all) of its images", "1410934": "", "1412500": "Thanks for sharing. This is helpful.", "1401318": "Thanks for sharing @hengck23 " } }