An Optimized Rate Control for Low-delay H.265/HEVC.

Area of Science:

• Video Compression
• Digital Signal Processing
• Information Theory

Background:

• High Efficiency Video Coding (H.265/HEVC) requires efficient rate control for low-delay applications.
• Traditional rate-distortion (R-D) models often rely on explicit parameter estimation, which can limit accuracy.
• Optimizing rate allocation at the Coding Tree Unit (CTU) level is crucial for effective HEVC compression.

Purpose of the Study:

• To propose an optimized CTU-level rate control approach for H.265/HEVC.
• To enhance rate control by incorporating the distributions of R-D model parameters, moving beyond traditional explicit estimation.
• To minimize the total expected distortion across CTUs within a defined bitrate budget.

Main Methods:

• Formulating CTU-level rate allocation as an optimization problem to minimize expected distortion.
• Considering the distributions of parameters within the rate-distortion model.
• Developing and applying a two-stage Bisection-based method to solve the optimization problem for the H.265/HEVC rate control algorithm.

Main Results:

• The proposed method demonstrates improved rate-distortion (R-D) performance.
• Quantifiable advances are shown through Bjøntegaard delta bit rate (BD-BR) metrics.
• The approach effectively minimizes expected distortion under rate constraints.

Conclusions:

• The optimized CTU-level rate control approach offers significant R-D performance gains for H.265/HEVC.
• Considering parameter distributions in R-D models enhances video compression efficiency.
• The Bisection-based method provides an effective solution for the proposed optimization problem in low-delay HEVC.