A python based MRI reconstruction toolbox with compressed sensing, parallel imaging and machine-learning functions
- "bloch_sim/" contains functions for MRI sequence simulation, these functions are designed for MR fingerprinting experiment
- "fft/" this is a wrap of FFT functions, i.e. cuFFT, FFTW, and NUFFT, implemented for both CPU and GPU
- "pics/" contains optimization algorithms, such as ADMM, conjugate gradient, gradient descent, for MRI compressed sensing and parallel imaging reconstructions, as well as operators such as total variation, Hankel matrix, coil sensitivity
- "neural_network/" contains a wrap of tensorflow functions for creating and testing neural_network, and zoo/ contains examples for full connection net, CNN, Unet, and FCN.
- "test/" contains testing code for above functions and something I am working on right now, e.g. MRI PICS reconstruction, IDEAL + CS reconstruction, FC or CNN for MRF quantification, Unet for creating mask on medical images
- Figure 1 MRIPY toolbox contains three major blocks: synthetic MRI, iterative/non-iterative reconstruction, and machine learning interface.
|
|
|