Methodologies for the Adaptive Compression of Video Sequences

In this work, a procedure for the video compression and transmission is presented, based on a Singular Value Decomposition approach, whose controls are obtained as the output of a constrained optimization problem that refers to the compression ratio as the optimization functional and an image quality index as the performance constraint. The tools of estimation theory allow to obtain a polynomial approximation of these indexes in a static fashion via least square technique, and adaptively, with the concurrent estimation of both the order of the polynomial functions and the salient function parameters through the use of Kalman Filters. The implementation of the whole system and some simulations of real video sequences are presented to validate and assess the proposed procedure.