Dynamic PET Image Reconstruction Using Spatiotemporal Kernel Method With Deep Image Prior

Prerequistites:

Python 3.7 (or 3.x)
PyTorch
Matlab R2021a

Neural KEM for dynamic PET reconstruction:

a). You can run 'demo.m' to test the proposed method.

b). 'DIP_process.py' is a function to run the step of DIP processing.

Required packages for PET reconstruction:

a). To use this package, you need to add the KER_v0.11 package into your matlab path by running setup.m in matlab. KER_v0.11 package can be downloaded from:

https://wanglab.faculty.ucdavis.edu/code

b). To test the algorithms in the package, run "demo.m" in the current folder. You may need your own system matrix G or use Jeff Fessler's IRT matlab toolbox to generate one. IRT can downloaded from

http://web.eecs.umich.edu/~fessler/code/index.html