Adaptive motion compensated vector quantizer for ATM video coding with cell loss compensation

Abstract

In this paper, a hue-layer adaptive motion compensated vector quantizer (AMCVQ) is proposed for ATM video coding. The proposed method classifies image blocks into three different motion classes. Low motion, medium motion and high motion. Only the motion compensated prediction (MCP) error of the medium-motion class and the pixel elements of the high-motion class are to be encoded by specific vector quantizers. Simulation results show that the proposed AMCVQ coder has better MSE performance than the MPEG DCT interframe coder. In addition, a simple compensation method is developed for handling the cell loss problem in ATM environment as well. >