Please use this identifier to cite or link to this item:
Title: MRI image reconstruction approach for partial K-space based on the Discrete Fourier Transform in the least square sense
Authors: Tobaji, Kalim
Matar, Laura
Abche, Antoine 
Karam, Elie 
Affiliations: Department of Electrical Engineering 
Department of Electrical Engineering 
Keywords: Medical image processing
Biomedical MRI
Conjugate gradient methods
Convex programming
Discrete Fourier transforms
Image reconstruction
Least squares approximations
Issue Date: 2015
Publisher: IEEE
Part of: Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA)
Start page: 44
End page: 49
Conference: Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA) (23-25 Sept. 2015 : Poznan, Poland) 
Magnetic Resonance Imaging (MRI) provides important and valuable information (images) about the organs structures and soft tissues non-invasively. In this work, a reconstruction approach using partial k-space, the "Optimized Discrete Fourier Transform" (ODFT), is introduced. The developed approach decomposes the 2-D Fourier Transform (FT) into two steps of 1-D DFT. The corresponding elements along each row or column are estimated using an optimization technique, namely, the complex conjugate gradient. It is proposed to be implemented in conjunction with other techniques to obtain an optimum image that is closer to the original image. The algorithm is evaluated visually and quantitatively using the Performance Test and the Mean Square Error as similarity measures. Also, its effectiveness is compared with different MRI reconstruction techniques such as the Projection onto Convex Set technique, the Conjugate Synthesis technique and the Zero filling technique. The results illustrate that the proposed technique outperforms the conventional MRI image reconstruction techniques.
Ezproxy URL: Link to full text
Type: Conference Paper
Appears in Collections:Department of Electrical Engineering

Show full item record

Record view(s)

checked on Jul 24, 2021

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.