search

ginobilinie / medSynthesisV1

This is a copy of package for medical image synthesis work with LRes-ResUnet and GAN (wgan-gp) in pytorch framework
MIT License
177 stars 44 forks source link

Data preprocessing #12

Closed luciaL closed 3 years ago

luciaL commented 4 years ago

Dear Dr. Nie Thank you for your great job. I'm working on medical image synthesis too. But I am new. I don't how to do preprocessing for CT values. Should I normalize it to 0-255 or just let be original
CT value. Look forward to your reply.

lucia

ginobilinie commented 4 years ago

@luciaL Thanks for your interest.

I think you'd better use some global constant (I mean statistics from the training set), for example, global max, global min, global mean, to convert the CT value to [-1, 1] or [-3, 3] or [0,1]

Then, you can recover the CT value to the normal region during testing stage using these global constants. CT intensities have physical meaning.

luciaL commented 4 years ago

@ginobilinie Thank you so much. It helps a lot.