![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F864495%2Ff4bc451b8425f8f0a991b94cb0c7fb78%2Ftable_4.png?generation=1568982234656800&alt=media\")
Qure.ai published in 2018 a very good related paper (same exact classes) and they even shared publicly their validation dataset with 500 strongly labeled samples. With their rigorous methodology, a variation of their paper was eventually published in one of the top medical journal The Lancet. I must admit that I admire this kind of public dissemination from a private startup.
\n\nPaper: https://arxiv.org/pdf/1803.05854.pdf\nDataset: http://headctstudy.qure.ai/dataset
\n\nThat is an excellent comparable baseline for this competition:\n
Qure.ai published in 2018 a very good related paper (same exact classes) and they even shared publicly their validation dataset with 500 strongly labeled samples. With their rigorous methodology, a variation of their paper was eventually published in one of the top medical journal The Lancet. I must admit that I admire this kind of public dissemination from a private startup.
\n\nPaper: https://arxiv.org/pdf/1803.05854.pdf\nDataset: http://headctstudy.qure.ai/dataset
\n\nThat is an excellent comparable baseline for this competition:\n
Did anyone get this to work, and if so can you comment on it?
\n\nThanks.
", "rawMarkdown": "Did anyone get this to work, and if so can you comment on it?\n\nThanks.", "replies": [ { "id": 655567, "postDate": "2019-10-23T07:41:59.863Z", "rawMarkdown": "" } ] }, { "id": 635534, "postDate": "2019-09-27T19:12:08.150Z", "content": "This is my first real competition while trying to take teh fast.ai courses. I spent all morning having problems getting the files unzipped in Paperspace even with 500gig. Took these over and it worked beautifully in Paperspace
", "rawMarkdown": "This is my first real competition while trying to take teh fast.ai courses. I spent all morning having problems getting the files unzipped in Paperspace even with 500gig. Took these over and it worked beautifully in Paperspace" }, { "id": 632741, "postDate": "2019-09-24T02:12:53.470Z", "content": "Has anyone been able to read the DICOM pixel data? I get various errors about pydicom having issues with it - something about a compressed format. It might also just be corrupt.
", "rawMarkdown": "Has anyone been able to read the DICOM pixel data? I get various errors about pydicom having issues with it - something about a compressed format. It might also just be corrupt." }, { "id": 630969, "postDate": "2019-09-21T04:59:41.293Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 644378, "author_name": "Bo Peng", "author_url": "", "post_date": "2019-10-08T18:00:53.717000", "content": "Did anyone get this to work, and if so can you comment on it?
\n\nThanks.
", "votes": 0, "replies": [ { "id": 655567, "author_name": "Tian Bingyang", "author_url": "", "post_date": "2019-10-23T07:41:59.863000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 635534, "author_name": "faheld", "author_url": "", "post_date": "2019-09-27T19:12:08.150000", "content": "This is my first real competition while trying to take teh fast.ai courses. I spent all morning having problems getting the files unzipped in Paperspace even with 500gig. Took these over and it worked beautifully in Paperspace
", "votes": 0, "replies": [] }, { "id": 632741, "author_name": "Tom H.", "author_url": "", "post_date": "2019-09-24T02:12:53.470000", "content": "Has anyone been able to read the DICOM pixel data? I get various errors about pydicom having issues with it - something about a compressed format. It might also just be corrupt.
", "votes": 0, "replies": [] }, { "id": 630969, "author_name": "", "author_url": "", "post_date": "2019-09-21T04:59:41.293000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "630582": "Qure.ai published in 2018 a very good related paper (same exact classes) and they even shared publicly their validation dataset with 500 strongly labeled samples. With their rigorous methodology, a variation of their paper was eventually published in one of the top medical journal The Lancet. I must admit that I admire this kind of public dissemination from a private startup.\n\nPaper: https://arxiv.org/pdf/1803.05854.pdf\nDataset: http://headctstudy.qure.ai/dataset\n\nThat is an excellent comparable baseline for this competition:\n\n\n\n", "644378": "Did anyone get this to work, and if so can you comment on it?\n\nThanks.", "635534": "This is my first real competition while trying to take teh fast.ai courses. I spent all morning having problems getting the files unzipped in Paperspace even with 500gig. Took these over and it worked beautifully in Paperspace", "632741": "Has anyone been able to read the DICOM pixel data? I get various errors about pydicom having issues with it - something about a compressed format. It might also just be corrupt.", "630969": "" } }