![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4422938%2F645cb8b2efccf16d0cc36406e88badf5%2Fimg900.png?generation=1590185906671194&alt=media\")
Hi Guys,
\n\nI have been analysing the data and I came across the fact that the images that are supposed to contain no tumour are showing cancer in the masks. Here are few examples
\n\nimage_id eef409493827bfb3b3d0e74e156923d4\ndata_provider radboud\nisup_grade 0\ngleason_score negative\nName: 9857, dtype: object
\n\n
image_id 007433133235efc27a39f11df6940829\ndata_provider radboud\nisup_grade 0\ngleason_score negative\nName: 18, dtype: object\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4422938%2Fa30f0176da69a7b475d650c98263eb86%2Fimage0.png?generation=1590186045487173&alt=media\")
I checked the first case you showed, and the mask for that image shows benign epithelium. Note that pixels labelled 2 are benign epithelium, and all labels >=3 are cancer.
", "rawMarkdown": "I checked the first case you showed, and the mask for that image shows benign epithelium. Note that pixels labelled 2 are benign epithelium, and all labels >=3 are cancer. ", "votes": 3 }, { "id": 857760, "postDate": "2020-05-22T22:20:49.037Z", "content": "Hi Guys,
\n\nI have been analysing the data and I came across the fact that the images that are supposed to contain no tumour are showing cancer in the masks. Here are few examples
\n\nimage_id eef409493827bfb3b3d0e74e156923d4\ndata_provider radboud\nisup_grade 0\ngleason_score negative\nName: 9857, dtype: object
\n\n
image_id 007433133235efc27a39f11df6940829\ndata_provider radboud\nisup_grade 0\ngleason_score negative\nName: 18, dtype: object\n
The masks for Radboud were generated by a deep learning algorithm, so this could be a false positive.
", "rawMarkdown": "The masks for Radboud were generated by a deep learning algorithm, so this could be a false positive. ", "votes": 1 }, { "id": 901445, "postDate": "2020-06-25T13:44:17.310Z", "content": "sometimes in the radboud masks, we have values more than 5. what should we do with that? can they be considered as 5?
", "rawMarkdown": "sometimes in the radboud masks, we have values more than 5. what should we do with that? can they be considered as 5?", "replies": [ { "id": 901477, "postDate": "2020-06-25T14:08:34.160Z", "content": "Could you give an example of a file that has values > 5?
", "rawMarkdown": "Could you give an example of a file that has values > 5? " } ] }, { "id": 862449, "postDate": "2020-05-26T15:39:31.530Z", "content": "Thanks for the responses. I will look into this
", "rawMarkdown": "Thanks for the responses. I will look into this" } ], "comments": [ { "id": 860777, "author_name": "Wouter Bulten", "author_url": "", "post_date": "2020-05-25T15:27:00.453000", "content": "I checked the first case you showed, and the mask for that image shows benign epithelium. Note that pixels labelled 2 are benign epithelium, and all labels >=3 are cancer.
", "votes": 3, "replies": [] }, { "id": 860641, "author_name": "Ian Pan", "author_url": "", "post_date": "2020-05-25T13:41:13.243000", "content": "The masks for Radboud were generated by a deep learning algorithm, so this could be a false positive.
", "votes": 1, "replies": [] }, { "id": 901445, "author_name": "aliechoes", "author_url": "", "post_date": "2020-06-25T13:44:17.310000", "content": "sometimes in the radboud masks, we have values more than 5. what should we do with that? can they be considered as 5?
", "votes": 0, "replies": [ { "id": 901477, "author_name": "Wouter Bulten", "author_url": "", "post_date": "2020-06-25T14:08:34.160000", "content": "Could you give an example of a file that has values > 5?
", "votes": 0, "replies": [] } ] }, { "id": 862449, "author_name": "Doctor_Q", "author_url": "", "post_date": "2020-05-26T15:39:31.530000", "content": "Thanks for the responses. I will look into this
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "860777": "I checked the first case you showed, and the mask for that image shows benign epithelium. Note that pixels labelled 2 are benign epithelium, and all labels >=3 are cancer. ", "857760": "Hi Guys,\n\nI have been analysing the data and I came across the fact that the images that are supposed to contain no tumour are showing cancer in the masks. Here are few examples\n\nimage_id eef409493827bfb3b3d0e74e156923d4\ndata_provider radboud\nisup_grade 0\ngleason_score negative\nName: 9857, dtype: object\n\n\n\nimage_id 007433133235efc27a39f11df6940829\ndata_provider radboud\nisup_grade 0\ngleason_score negative\nName: 18, dtype: object\n\n", "860641": "The masks for Radboud were generated by a deep learning algorithm, so this could be a false positive. ", "901445": "sometimes in the radboud masks, we have values more than 5. what should we do with that? can they be considered as 5?", "862449": "Thanks for the responses. I will look into this" } }