Frequency components selected based on gravitational search algorithm for magnetic particle imaging reconstruction

Abstract

 It often utilizes system matrix to reconstruct in Magnetic Particle Imaging (MPI), but it is time-consuming and memory-intensive. Therefore, use Signal-to-Noise Ratio (SNR) to reduce frequency components to speed up and reduce memory, but only use SNR values do not contain other information about the frequency and may lose crucial information required for reconstruction. To address this limitation, the frequency components selection based on gravitational search algorithm (GSA) method is proposed herein, this method leverages Newton's Law of Universal Gravitation and the Law of Kinematics to intelligently select frequency components, potentially enhancing the reconstruction image quality in MPI. Experimental results demonstrate favorable quantitative indices for the reconstructed images.