sigpy is a Python package for signal reconstruction, with GPU support using cupy.
sigpy provides commonly used signal processing functions, including convolution, FFT, NUFFT, wavelet transform, and thresholding functions. All operations, except wavelet transform, can run on GPU. These operations are wrapped in a linear operator class (Linop) or a proximal operator class (Prox) for easy usage in iterative algorithms. sigpy also implements commonly used iterative algorithms, such as conjugate gradient, (accelerated/proximal) gradient method, and primal dual hybrid gradient.
sigpy provides a submodule sigpy.mri that uses the core module to implement common MRI iterative reconstruction methods, including SENSE reconstruction, l1-wavelet reconstruction, total-variation reconstruction, and JSENSE reconstruction. In addition, it provides convenient simulation and sampling functions, such as poisson-disc sampling function.
sigpy also provides a preliminary submodule sigpy.learn that implements convolutional sparse coding, and linear regression.
The package is on PyPI, and can be installed via pip:
pip install sigpy
For optional gpu support, the package requires cupy.
For optional distributed programming support, the package requires mpi4py.
Alternatively, the package can be installed from source with the following requirements:
- python3
- numpy
- scipy
- pywavelets
- numba
Our documentation is hosted on Read the Docs: https://sigpy.readthedocs.io