Thesis (D.Eng.)--Chulalongkorn University, 2007

Many multimedia applications require video coding schemes that can provide an acceptable quality of service to the end users. Due to the bandwidth of channel are limited, video source is usually encoded with a low-bit-rate video compression. In the coding process, the mean bit rate of the coded data is proportional to the decoded video quality. The process to allocate suitable number of bits to each video frame is called rate control. To achieve best video quality subject to the bit rate constraints, rate control algorithm has to be efficiently designed with the underlying rate-distortion model to cope with several scenarios especially low bit-rate and low delay scenario which are very important factors for real time video communication. In this dissertation, we verified through analysis and simulations that cauchy distribution provides more accurate estimates of rate and distortion characteristics of video sequences than the previously distribution used such as Laplacian distribution. A new rate-distortion optimization model based on cauchy probability density function is then proposed and optimized to find the optimum choice of quantization step size in video coding by using Lagrange multiplier technique. The proposed rate-distortion optimization model is then applied in an application of rate control. Thus, we propose a new rate control for low bit-rate H.264 video coding. Considering the scenario of low delay in real time video transmission that frames may be skipped to reduce the buffer delay. This results in the motion discontinuity and low visual quality in video communication system. To alleviate the problem, we propose another rate control scheme for low bit-rate H.264 video coding under low delay constraint. Simulation results for both proposed rate control schemes indicate video quality improvement with less number of frame skipped compared to H.264 JM 8.6