Contrast optimization technique for multi-planar magnetic resonance images of the brain

A magnetic resonance (MR) image processing technique is presented which attempts to improve brain tissue contrast on sagittal views of the brain using the higher contrast definition of the same tissue from corresponding axial planes. The method uses median filtering in order to maximize the contrast-to-noise ratio in axial images, and subsequently transforms sagittal views using a simplified histogram matching algorithm suitable for real-time contrast correction using PC-based imaging workstations. The transformation has been tested on a number of multiplanar MR studies of the brain, yielding acceptable results for overall contrast enhancement in sagittal views. In the extreme case, and especially when the gray palette available for the display is relatively narrow, the expansion of the integer image data from a narrow range into a wider one causes a salt-and-pepper effect in the brain area. The intensity values of white and gray matter are forced apart, but they coalesce around discrete steps in intensity. As a result, upon these conditions the image appears noisier. A low-pass filter or median filter-based smoother can be applied to soften the effect at the expense of lower overall resolution in the image.