A Moment Based Fractional Fourier Transform Scheme for MR Image Classification
Stamatis Mastromichalakis, Spiros Chountasis · Automatic Control and Computer Sciences · 2021
Abstract The classification of tomography images into unhealthy or healthy is a key preclinical process within medical image processing. The training and feature extraction comprise the most time and memory consuming processes in a classifier. This paper introduces a scheme that employs fractional Fourier transform and Fourier moment analysis for feature extraction. The fractional Fourier transform was used to obtain the unified time–frequency spectrum of images, and this, combined with Fourier moments analysis, seems to can improve the operation of classifying a given brain image as normal or abnormal. The extracted features of this procedure were reduced by principal component analysis and then are used as input vectors to the support vector machine algorithm. To demonstrate the potential of the technique presented in this paper in comparison with other state-of-the-art methods, numerical simulations are performed.