Generalized Nash Bargaining Solution to Rate Control Optimization for Spatial Scalable Video Coding.

Area of Science:

• Computer Science
• Electrical Engineering
• Information Theory

Background:

• Rate control (RC) is crucial for scalable video coding (SVC) in storage and streaming.
• Inner-layer bit allocation in SVC resembles a bargaining problem, suggesting game theory applications.

Purpose of the Study:

• To propose a bargaining game-based one-pass RC scheme for spatial H.264/SVC.
• To enhance RC performance for spatial SVC through optimal bit allocation.

Main Methods:

• Modeling encoding constraints (bit rates, buffer size) as resources in an inner-layer bit allocation bargaining game.
• Utilizing a modified rate-distortion (R-D) model with inter-layer coding information.
• Applying the generalized Nash bargaining solution (NBS) for optimal bit allocation and adaptive bandwidth assignment based on bargaining powers.

Main Results:

• The proposed RC algorithm significantly improves image quality.
• Demonstrates enhanced buffer smoothness during video streaming.
• Achieves an average mismatch within 0.19%-2.63%.

Conclusions:

• The bargaining game framework effectively optimizes bit allocation in spatial SVC.
• The proposed method offers a robust solution for rate control in H.264/SVC.
• Results indicate superior performance in terms of quality and buffer management.