Simulation of the effect of bit assignment to multi-hypothesis pictures per group for motion compensated video compression (MCVC)

Abstract

This paper describes the Lagrange based formulation of bit-assignment for multi-hypothesis frames over number of pictures in a group for motion compensated video compression. Theoretical insight on bit assignment is accomplished by applying Lagrange multiplier and its cost function for a real-time computational analysis. Video signals are compressed at optimal bit allocation for qualitative reconstruction at the decoder (receiver) depending upon the number of hypothesis used in coding the signal. This has made multi-hypothesis motion compensation (MHMC) a great process in the prediction of the actual signal through the combination of more than one motion compensated prediction (MCP). Optimal Bits were allocated to multi-hypothesis frames using MATLAB to optimize the number of hypothesis frames per group of pictures (GOP) decoded for an improved signal compression. Simulation results show that an optimal bit assignment of the range 0.089 bpp to 0.159 bpp was realized with eight numbers of hypotheses frames per (GOP). This implies that varying number of hypotheses actualizes optimal bit assignment for motion compensation. This shows a qualitative signal compression performance analysis for high technology motion pictures in communication services and storage processes. It also addresses a better prediction due to greater time correlation between hypotheses and helps in the improvement of coding and reconstruction quality of video signals.

Key words: Bit-rate, frame, Group of Pictures (GOP), Motion Compensation,multi-hypothesis, Video Compression.