Image Reconstruction Scheme Based on Phase Correction and Singularity Function Analysis Model

Abstract

The singularity function analysis (SFA) model is a mathematical tool that allows representing an image with much less coefficients than the conventional Fourier transform (FT) model while maintaining a better image quality. The performance of the SFA method for reconstructing medical images may however degrade if unwanted phase drifts are present in the acquired image. This paper proposes a new SFA-based reconstruction scheme by taking into account phase drifts. To this end, phase drifts are first mathematically formulated and corrected. The singularity function model is then applied to represent the phase-corrected image. The performance of this phase-corrected SFA reconstruction scheme is evaluated using both simulated and real brain images, and compared with the conventional FT model and SFA method without phase correction. The results demonstrate that the proposed reconstruction method achieves significant improvement in image quality.