I was trying to convert the single dicom channel for medical imaging to 3 channels so I could use a pre-trained resnet to work on it (for kaggle competition, but more to learn). However I am having trouble with the built in code. Does any body have some suggestion ? Would it be better to try a different pre-trained model ? Or can we more easily convert the 1 channel to 3 channels.
Cheers 
p.s. this is my first topic so apologies if I did something wrong 