Is there a way to distinguish the source of the data?

Hi orgnizers, big thanks for the contributions! I was downloading the data, and I wonder is there a way to distinguish the source of the data, i.e., can we know which images are from JSRT and which are from PadChest?

Hi!

Thanks for your question. We provide the original names of the files we have selected (as well as the data source they come from ) in cxr_images/original_data/filenames_orig_and_new.csv.

Hope this answers your question.

Best, Ecem

Many thanks to the quick reply! I shall check the names and try to relate the ids back to their original domains after the downloading.

If I am right this file only covers the nodule data. However, is there name mapping information regarding the "healthy" subjects? best Finn

Hi Finn,

Thanks for noticing this. Yes, i have just uploaded the csv file for non-nodule image mapping here: https://drive.google.com/file/d/1CiO-d02vhpKivAYJqe9FS6iz7rd5PQ9I/view?usp=sharing

I will also upload this to zenodo soon.

Best, Ecem