Help me
fsy33 opened this issue · 5 comments
When I run your project, the data path is displayed incorrectly. I would like to ask how you process the original data set into the index form used in the project. If it is convenient, can you provide the preprocessing code of the data set? Thank you!
You can see in the /data/fold{}.txt files that the indexing has already been done, so you only need to modify the image paths accordingly. You would need to download the IU X-ray images first, and modify the image paths in /data/fold{0,1,2,3,4}.txt to your image directory. Hope this helps!
Thank you for your reply. When your model reads the index file, there are only two pictures in one line of the index file, one on the front and one on the side. However, in the original data set, there are several pictures in the corresponding chest film folder of some patients. I wonder whether I can read all the pictures in that corresponding folder every time. I wonder if this will affect the subsequent data processing. Do you think this method is feasible? I look forward to your reply.
In our work, we only read one frontal-view image and one lateral-view image for each study to make it comparable with previous baseline methods. You can read all images in each folder every time, but this was not currently supported by our codebase.
Thank you for your reply, which is very helpful to me!
Thank you for your reply, which is very helpful to me!