Jensen-renyi divergence measure: theoretical and computational perspectives

We analyze the theoretical properties of the Jensen-RBnyi divergence, and we discuss its appli- cation to parametric classification. Numerical results on medical images show the the potential of the pro- posed divergence measure in image registration. I. THEORETICAL AND COMPUTATIONAL ANALYSIS Definition 1 Let p,,p,, . . . ,p, be n probabzlzty dzstrzbutzons of X and w = (w1, w2,. . . ,wn) be a wezght vector such that C:=lwz = 1 and wz 2 0. The Jensen-Rknyi dzvergence zs defined as n JRf(p,, . . . ,pn) =-Re (CWiPi) - xwzRe(Pi)> i=l i=l

[1]  Paul A. Viola,et al.  Alignment by Maximization of Mutual Information , 1997, International Journal of Computer Vision.

[2]  Yun He,et al.  A generalized divergence measure for robust image registration , 2003, IEEE Trans. Signal Process..