Contrast enhancement of mr brain images by canonical correlations based kernel independent component analysis

Since the MR signals can be considered as a combination of the signals from each brain matters, it has been shown that independent component analysis (ICA) can be used for contrast enhancement of MR images. However, ICA is a linear method in nature, and it is inadequate to well-describe nonlinear variations of the real MR images. In this paper, we propose a new method for contrast enhancement of MR brain images using a canonical correlation based kernel independent component analysis (KICA). Experimental results on both phantom MR datasets and real clinical MR datasets show that the contrast of MR images can be significantly enhanced by KICA.