A framework of MRI fat suppress imaging fusion system: MRI inhomogeneity correction for femur abnormality analysis

Abstract

Short Inversion Time Inversion Recovery (STIR) is one of the fat suppression imaging techniques which involves inversion recovery pulse sequences. STIR uses time delay TI to nullify longitudinal magnetisation. This is to suppress the fat signals from tissues and increase the echo time (TE) so that another sequence called T2 will increase the contrast between tissues. STIR sequence has greater advantage of display signal based on water information about a patient. However, intensity inhomogeneity is a common artifact occurs in MRI machines due to radio-frequency non-uniformity and static field-inhomogeneity. The purpose of this study is to propose a method uses image fusion system, imitating the MRI machine to produce STIR or any other fat suppressed images. The image fusion system could completely remove the fat signals and correct inhomogeneity which occurs in the MRI machine. Image quality measurement on Mean Square Error (MSE) is low (0.055069) and Similarity Index (SSIM) is high (0.995606) for 4 clusters. This show that the resultant fused `STIR' like image has high similarity with the original MRI STIR.