Estimating the value of θ in the intra frame for ρ-domain rate control algorithms

Authors:
Zhihang Wang;Shengfu Dong;Long Xu;Wen Gao;Qingming Huang
Affiliations:
Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China;Institute of Digital Media, Peking University, Beijing, China;Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China;Institute of Computing Technology, Chinese Academy of Sciences and Institute of Digital Media, Peking University, Beijing, China;Graduate School, Chinese Academy of Sciences, Beijing, China
Venue:
PCS'09 Proceedings of the 27th conference on Picture Coding Symposium
Year:
2009

Citing 7
Cited 0

Context-based 2D-VLC entropy coder in AVS video coding standard

Journal of Computer Science and Technology - Special section on China AVS standard
A mathematical analysis of the DCT coefficient distributions for images

IEEE Transactions on Image Processing
Rate control of MPEG video coding and recording by rate-quantization modeling

IEEE Transactions on Circuits and Systems for Video Technology
A new rate control scheme using quadratic rate distortion model

IEEE Transactions on Circuits and Systems for Video Technology
Bit-rate control using piecewise approximated rate-distortion characteristics

IEEE Transactions on Circuits and Systems for Video Technology
Rate control in DCT video coding for low-delay communications

IEEE Transactions on Circuits and Systems for Video Technology
Optimum bit allocation and accurate rate control for video coding via ρ-domain source modeling

IEEE Transactions on Circuits and Systems for Video Technology

Quantified Score

Hi-index	0.02

Visualization

Abstract

The ρ-domain rate control algorithm is very simple and effective. There is only one parameter, θ, used in the model. θ reflects the relationship between the rate and the quantization step. Accurately estimating the θ value of the I (intra) frame is important for an efficient ρ-domain rate control algorithm because the I frame will influence the remaining frames in the GOP (group of pictures) greatly. In this paper, we propose to estimate the θ value based on the energy of the frame. The energy difference of the I frames is used to predict the new θ. A linear relationship between θ and the energy is deduced. Experimental results show that using the energy can more accurately predict the θ value of I frame than only using the θ value of the previous P (predict) frame.