Parameter space dimension reduction for multidimensional signals differential pulse-code modulation interpolator

Abstract

In this paper, new parameterized differential pulse-code modulation interpolator is proposed. It uses the correspondence of signal samples in the processed sample vicinity to select the interpolation method. Interpolation error of proposed interpolator is minimized during the optimization procedure. Before the optimization, the parameter range dimension is reduced in order to lower the computational complexity. Experimental research shows that proposed interpolator outperforms the non-parameterized one