search

uci-cbcl / NoduleNet

[MICCAI' 19] NoduleNet: Decoupled False Positive Reduction for Pulmonary Nodule Detection and Segmentation
Other
185 stars 57 forks source link

dataset details #13

Closed czifan closed 4 years ago

czifan commented 4 years ago

Thanks for the open source code, which is so helpful for me! But I have a question about the dataset. I found that "resulting in a total number of 586 CT scans with 1131 nodules" in section "3 Result - Data and experiment configurations" of your paper. But I noticed that there are only 583 CT scans in the split folder in this code. In addition, I have preprocessed the data by following this readme.md and I only have 583 CT scans. Is the results in the paper on this dataset of this code? Because it will be a baseline for comparison and this question is important. Thank you!!!

tanghaotommy commented 4 years ago

Thanks for your interest.

It is the code for this paper. I took a look and it seems the code indeed generates 583 CTs. I will look into this problem and figure out which number is correct.

czifan commented 4 years ago

Thanks for your reply! Could you tell me the result of the problem if you solve it?

huliangxiao commented 4 years ago

Thank you so much for the code! I have the same problem as @czifan, please let us know when it's settled.

tanghaotommy commented 4 years ago

Hello, I just double checked. There are a total of 583 CT scans that meet the criteria, however, just for sanity check, I added 3 more CT scans randomly (so a total of 586). But for evaluation, only the 583 CT scans were valid

czifan commented 4 years ago

Thank you very much!!!