In this paper we propose a novel framework for the multi-objective optimization of a video CODEC based on genetic algorithms. The proposed framework is designed to jointly minimize the complexity, memory usage (both at the encoder and decoder), bit rate and to maximize the quality of the compressed video stream. In particular, in our present attempt the optimization strategy is designed to determine the optimum coding parameters for a H.264 AVC video codec in a memory and bandwidth constrained environment. This is demonstrated through extensive experiments and mathematical formulation that results in the optimum solution/s to the multi-objective optimisation problem being found. We show that such an approach is highly desirable in obtaining optimum coding parameters for video delivery over the internet, where a feedback channel from the decoder to the encoder is practical. .
This chapter presents a generalised framework for multi-objective optimisation of video CODECs for use in off-line, on-demand applications. In particular, an optimization scheme is proposed to determine the optimum coding parameters for a H.264 AVC video codec in a memory and bandwidth constrained environment, which minimises codec complexity and video distortion. The encoding/decoding parameters that have a significant impact on the performance of the codec are initially obtained through experimental analysis. A mathematical formulation by means of regression is subsequently used to associate these parameters with the relevant objectives and define a Multi-Objective Optimization (MOO) problem. Solutions to the optimization problem are reached through a Non-dominated Sorting Genetic Algorithm (NSGA-II). It is shown that the proposed framework is flexible on the number of objectives that can jointly be optimized. Furthermore, any of the objectives can be included as constraints depending on the requirements of the services to be supported. Practical use of the proposed framework is described using a case study that involves video content transmission to a mobile hand.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.