I am puzzled by the logic in the example pre-processing scripts provided at https://github.com/aur-dream/dm-caffe-example. dm-caffe-preprocessing-example/generate_labels.py is labeling **both** left and right views as cancerous independent of which breast (left or right) developed cancer within 12 months of the exam. Is that appropriate given that both sub-challenge 1 and 2 require predictions specific to each breast?

Created by Prasad Chodavarapu chprasad
Thanks for your question, Prasad. Thanks also for picking up on this logic error. This logic was based on the challenge's previous specification, which was to make predictions at the subject level (now deprecated). We have now uploaded a corrected version of _generate_labels.py_ here: https://github.com/aur-dream/dm-caffe-example/blob/master/dm-caffe-preprocessing-example/generate_labels.py. Thanks and best regards, Shivy

Logic in example pre-processing scripts page is loading…