Research On The Rate Control Algorithm For Scalable Video Coding

Scalable Video Coding is currently a very active working area in the research community and international standardization, which is a highly attractive solution to the problems of today's new video service requirements. In this paper, we conducted a study and analysis of the rate control algorithm for the Scalable Video Coding based on the applications of the Scalable Video Coding.Firstly, described the key technology to achieve temporal, spatial and quality scalability in the Scalable Video Coding, as well as the Cascaded QP assignment. Researched and analysis the rate control algorithms in the reference software JSVM of the Scalable Video Coding. This laid a foundation for the following study.Next, we presented aρ-domain model based rate control algorithm for the Scalable Video Coding. Hierarchical-B frame prediction structure has been introduced into Scalable video coding, it is a pyramid structure. Therefore, we improved the bit allocation method based on such pyramid structure. The different types of frames gave different contribution to the overall coding performance. Our bit allocation strategy was based on the complexity and weight factor of the frame. At the same time, we used the PID controller to adjust the initial distribution bits. Experiments illustrate that compared to the current rate control algorithm FixedQP in JSVM, the algorithm based on theρ-domain model for Scalable Video Coding proposed in this paper has better performance in terms of both the accuracy of rate matching and PSNR gain.Finally, we presented a rate-distortion model based rate control algorithm for the Scalable Video Coding. Compared to the Laplacian distribution model, the distribution of DCT transform coefficients is more in line with the Cauchy distribution. According to the simulation results, we applied Cauchy distribution based rate-distortion model for I and P frames, and linear rate-distortion model for the B frames. This classification rate-distortion model method achieved a bit-rate control performance and computational complexity trade-offs. Experiments illustrate that compared to the current rate control algorithm FixedQP in JSVM, our algorithm for Scalable Video Coding has better performance in terms of both the accuracy of rate matching and PSNR gain.