Main Article Content
This work is licensed under a Creative Commons Attribution 4.0 International License.
Magnetic particle imaging (MPI) is a rapidly developing medical imaging modality, which uses the nonlinear response of superparamagnetic iron oxide nanoparticles to the applied magnetic field to image their spatial distribution. Due to the direct feedthrough of excitation signals, the existing MPI systems directly filter out the fundamental frequency component of the received signal, resulting in the loss of first harmonic information. In this work, we proposed a deep learning (DL) method adopting self-attention mechanism, which can effectively recover fundamental frequency component of the signals in the presence of background noise. At the same time, our method deals with two-dimensional time-frequency spectrum obtaining by piecewise Fourier transform from the time domain signals. The performance of our method is analyzed via simulation experiments, which show that our method can effectively recover first harmonic information and obtain high quality MPI reconstructed images.