Brain segmentation in magnetic resonance images using fast fourier transform

Magnetic Resonance Images (MRI) are used to produce images of soft tissues of human body. It is used to analyze the human organs. Automatic detection and Segmentation of the brain can result in early detection and intervention for a number of brain diseases. In this paper, we propose Fast Fourier Transform (FFT) based method to remove or subdue the low intensity noise and thus it makes accurate segmentation of brain images. However, accurate segmentation of the MRI images is very important and crucial for the exact diagnosis by computer aided clinical tools. In this work, T2-weighted image is first transformed into frequency domain. By multiplying FFT with high pass filter, we obtained a filtered image. The inverse transform function is used to get the real part of the filtered image. By using thresholding technique, we removed low intensity from the filtered image. Assuming largest connected component is brain in MRI brain images, labeling and largest connected component (LCC) techniques are used to obtain the brain mask from which the brain is segmented.