A supervised neural constant bit rate video controller for MPEG2 encoders

Abstract

A supervised training algorithm is used in a radial basis function neural network in order to improve the performance of a recently introduced nonlinear predictive rate controller for MPEG2 encoders. The algorithm, which is based on the stochastic gradient method, is used for updating the radial basis centers. By means of extensive computer simulation with standard video sequences, practical design parameters are presented for buffer control in constant bit rate MPEG2 video encoders.