Trainable Fractional Fourier Transform

In this repository, we present the source code for the experiments of our Trainable Fractional Fourier Transform paper is accepted to IEEE Signal Processing Letters. The installable package torch-frft is maintained at its own GitHub page. The package is available on both PyPI and Conda. Installation instructions are provided below. Please use the following BibTeX entry to cite our work:

@article{trainable-frft-2024,
  author   = {Koç, Emirhan and Alikaşifoğlu, Tuna and Aras, Arda Can and Koç, Aykut},
  journal  = {IEEE Signal Processing Letters},
  title    = {Trainable Fractional Fourier Transform},
  year     = {2024},
  volume   = {31},
  number   = {},
  pages    = {751-755},
  keywords = {Vectors;Convolution;Training;Task analysis;Computational modeling;Time series analysis;Feature extraction;Machine learning;neural networks;FT;fractional FT;deep learning},
  doi      = {10.1109/LSP.2024.3372779}
}

Installation of torch-frft

You can install the package directly from PyPI using pip or poetry as follows:

pip install torch-frft

or

poetry add torch-frft

or directly from Conda:

conda install -c conda-forge torch-frft