Action-Constrained Reinforcement Learning for Frame-Level Bit Allocation in HEVC/H.265 through Frank-Wolfe Policy Optimization

TL;DR

This paper introduces a novel reinforcement learning framework using Frank-Wolfe policy optimization for frame-level bit allocation in HEVC/H.265, improving convergence and performance over existing methods.

Contribution

It proposes Neural Frank-Wolfe Policy Optimization (NFWPO), a new RL approach that guarantees convergence and enhances bit allocation efficiency in video encoding.

Findings

Outperforms single-critic and dual-critic RL methods in VMAF optimization.

Achieves comparable rate-distortion performance to x265's 2-pass control.

Demonstrates improved training stability and convergence guarantees.

Abstract

This paper presents a reinforcement learning (RL) framework that leverages Frank-Wolfe policy optimization to address frame-level bit allocation for HEVC/H.265. Most previous RL-based approaches adopt the single-critic design, which weights the rewards for distortion minimization and rate regularization by an empirically chosen hyper-parameter. More recently, the dual-critic design is proposed to update the actor network by alternating the rate and distortion critics. However, the convergence of training is not guaranteed. To address this issue, we introduce Neural Frank-Wolfe Policy Optimization (NFWPO) in formulating the frame-level bit allocation as an action-constrained RL problem. In this new framework, the rate critic serves to specify a feasible action set, and the distortion critic updates the actor network towards maximizing the reconstruction quality while conforming to the action constraint. Experimental results show that when trained to optimize the video multi-method assessment fusion (VMAF) metric, our NFWPO-based model outperforms both the single-critic and the dual-critic methods. It also demonstrates comparable rate-distortion performance to the 2-pass average bit rate control of x265.