Image Enhancement Technique for MRI Brain Images

Magnet Resonance Imaging (MRI) is a medical imaging technique that produces high-quality images of human body organs. MRI is mostly used to analyze brain images and diagnose brain tumors. MRI images suffer from noise and low varying contrast. The noise is the main factor that degrades the image quality and affects the image's contrast. The image enhancement technique is proposed, and it contains only three stages. The first stage handles the noise level of MRI brain images using a sub-space least square estimator, the second stage is for contrast normalization using the Retinex algorithm, and the third stage is for contrast enhancement using independent component analysis. The Oasis image database contains data from 38 patients and three sets of image views of each patient as axial, coronal, and sagittal images. Peak signal-to-noise ratio (PSNR) and contrast improvement ratio are evaluation parameters. It is observed that the axial and coronal image databases gave an efficient contrast improvement ratio of 1.4 at 2.5dB and 1.29 at 2.29dB, respectively, compared to the sagittal image view of 1.18 at 2.35dB.