1

A major problem of segmentation of magnetic res-onance imaging is that intensities are not standard-ized like in computed tomography. In this article we will present a new method for MRI intensity standardization by aligning histograms of higher di-mensions. So the correction process is independent from spatial coherences and prior segmentations of the reference and newly acquired images. Although the approach is not limited to a specic application or area of the body, it is utilized for fast classica-tion of brain tissue. Therefore, reference statistics are computed once using a hidden Markov random eld approach. The intensity corrected images are then classied using these learned statistics. In or-der to evaluate the presented methods the mean dif-ferences of intra-patient time series are chosen as reference. Furthermore, results of fast brain tissue classication are presented.