Title: FPGA hardware architecture of correlation-based MRI images classification using XSG

Authors: Fayçal Hamdaoui; Anis Sakly; Abdellatif Mtibaa

Addresses: Faculty of Sciences of Monastir, Laboratory of EµE, University of Monastir, Monastir, Tunisia ' Industrial Systems Study and Renewable Energy (ESIER), National Engineering School of Monastir (ENIM), Electrical Department, University of Monastir, Av Ibn ElJazzar 5019, Monastir, Tunisia ' Laboratory of EµE, Faculty of Sciences of Monastir, Electrical Department, National Engineering School of Monastir (ENIM), University of Monastir, Monastir, Tunisia

Abstract: Medical imaging classification is one of the areas where using algorithm-based hardware architecture improves performance, in terms of time processing. It gives better and clearer results than when using software implementation. Today, advantages of field-programmable gate array (FPGA), including reusability, filed reprogramability, simpler design cycle, fast marketing and a combination of the main advantages of ASICs and DSPs make them powerful and very attractive devices for rapid prototyping of all images processing applications. In this paper, we use Xilinx system generator (XSG) environment to develop a hardware classification-based correlation algorithm from a system level approach. This architecture may be of great influence on the final choice to prove if the MRI image is with lesions brain or normal. Results are illustrated on a simple example for brain magnetic resonance imaging (MRI) images classification. Two sets are used: a set of normal MR images and another set with MR lesion brain images.

Keywords: MRI classification; brain MRI; brain scans; magnetic resonance imaging; image correlation; field-programmable gate arrays; FPGA based hardware implementation; XSG; Xilinx system generator; MatLab; MRI images; image classification; medical imaging; rapid prototyping; image processing; brain images.

DOI: 10.1504/IJCAT.2015.071422

International Journal of Computer Applications in Technology, 2015 Vol.52 No.1, pp.77 - 85

Published online: 27 Aug 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article