Application of generalized Jensen-Schur measure in medical image registration
广义Jensen-Schur测度在医学图像配准中的应用

For the influences of noise, interpolation and image modality, the medical image registration method based on mutual information or normalized mutual information would cause local extrema, small convergence area, and even inaccurate registration. A new generalized Jensen-Schur measure was defined, which used "nonlinear increasing" of butterworth function to eliminate false extrema. Four new generalized Jensen-Schur measures, mutual information and normalized mutual information were analyzed and compared by applying them to rigid registration. The results of tests show that the new constructed JS22 and JS23 measures outperform other measures in noise immunity and convergence, and eliminating false extrema caused by PV interpolation.