AIM-Harvard / DeepCAC

Fully automatic coronary calcium risk assessment using Deep Learning.
GNU General Public License v3.0
37 stars 29 forks source link

Failed on Cardiac images #8

Open amirjaber opened 2 years ago

amirjaber commented 2 years ago

Hi, Thanks for sharing the model weights and the code. I ran the model sucessfully on 4 available test data. Upon running on my own cardiac dataset, it failed to localize the heart and then segment the heart. Reading your paper, I noticed that you have only trained with cardiac gated images and no non-gated chest CT was used for training. Pls let me know if I should change the input image size for my data to go through step 1 and 2? right now my image size is 512x512x50