The number of training label masks is not equal to the number of training images. Why is that so?
", "messages": [ { "id": 818050, "postDate": "2020-04-23T15:51:46.933Z", "content": "@venky2506 The organisers have mentioned this in the data description. You can take a look at it here. They don't really explain why though.
\n\"train label masks: Segmentation masks showing which parts of the image led to the ISUP grade. Not all training images have label masks, and there may be false positives or false negatives in the label masks for a variety of reasons. \"
", "rawMarkdown": "@venky2506 The organisers have mentioned this in the data description. You can take a look at it [here](https://www.kaggle.com/c/prostate-cancer-grade-assessment/data). They don't really explain why though.\n\n\"train label masks: Segmentation masks showing which parts of the image led to the ISUP grade. **Not all training images have label masks**, and there may be false positives or false negatives in the label masks for a variety of reasons. \"", "votes": 1, "replies": [ { "id": 846107, "postDate": "2020-05-13T15:19:42.470Z", "content": "To answer your question: We had some cases where our automated method failed to predict the label for the majority of the glandular tissue. Those masks were discarded but we left the images as part of the dataset.
", "rawMarkdown": "To answer your question: We had some cases where our automated method failed to predict the label for the majority of the glandular tissue. Those masks were discarded but we left the images as part of the dataset. ", "votes": 2 } ] }, { "id": 845963, "postDate": "2020-05-13T14:07:01.420Z", "content": "There are 10616 images and 10516 masks. I guess it is ok to work with \"just\" 10516 images hehe.
", "rawMarkdown": "There are 10616 images and 10516 masks. I guess it is ok to work with \"just\" 10516 images hehe. " }, { "id": 816203, "postDate": "2020-04-22T07:09:30.710Z", "content": "The number of training label masks is not equal to the number of training images. Why is that so?
", "rawMarkdown": "The number of training label masks is not equal to the number of training images. Why is that so?", "votes": -1 } ], "comments": [ { "id": 818050, "author_name": "Rai", "author_url": "", "post_date": "2020-04-23T15:51:46.933000", "content": "@venky2506 The organisers have mentioned this in the data description. You can take a look at it here. They don't really explain why though.
\n\"train label masks: Segmentation masks showing which parts of the image led to the ISUP grade. Not all training images have label masks, and there may be false positives or false negatives in the label masks for a variety of reasons. \"
", "votes": 1, "replies": [ { "id": 846107, "author_name": "Wouter Bulten", "author_url": "", "post_date": "2020-05-13T15:19:42.470000", "content": "To answer your question: We had some cases where our automated method failed to predict the label for the majority of the glandular tissue. Those masks were discarded but we left the images as part of the dataset.
", "votes": 2, "replies": [] } ] }, { "id": 845963, "author_name": "Leslie Solorzano", "author_url": "", "post_date": "2020-05-13T14:07:01.420000", "content": "There are 10616 images and 10516 masks. I guess it is ok to work with \"just\" 10516 images hehe.
", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "818050": "@venky2506 The organisers have mentioned this in the data description. You can take a look at it [here](https://www.kaggle.com/c/prostate-cancer-grade-assessment/data). They don't really explain why though.\n\n\"train label masks: Segmentation masks showing which parts of the image led to the ISUP grade. **Not all training images have label masks**, and there may be false positives or false negatives in the label masks for a variety of reasons. \"", "845963": "There are 10616 images and 10516 masks. I guess it is ok to work with \"just\" 10516 images hehe. ", "816203": "The number of training label masks is not equal to the number of training images. Why is that so?" } }