Advanced Learning-Based Inter Prediction for Future Video Coding

TL;DR

This paper introduces a low complexity neural network-based inter prediction method for AVS4 video coding, replacing traditional filters with learned parameters to improve efficiency and coding gain.

Contribution

It presents a lightweight neural network approach for inter prediction in AVS4, enabling faster inference and seamless integration into existing codecs.

Findings

Achieves up to 0.31% coding gain in U component.

Replaces traditional filters with learned parameters.

Enables faster inference without third-party dependencies.

Abstract

In the fourth generation Audio Video coding Standard (AVS4), the Inter Prediction Filter (INTERPF) reduces discontinuities between prediction and adjacent reconstructed pixels in inter prediction. The paper proposes a low complexity learning-based inter prediction (LLIP) method to replace the traditional INTERPF. LLIP enhances the filtering process by leveraging a lightweight neural network model, where parameters can be exported for efficient inference. Specifically, we extract pixels and coordinates utilized by the traditional INTERPF to form the training dataset. Subsequently, we export the weights and biases of the trained neural network model and implement the inference process without any third-party dependency, enabling seamless integration into video codec without relying on Libtorch, thus achieving faster inference speed. Ultimately, we replace the traditional handcraft filtering parameters in INTERPF with the learned optimal filtering parameters. This practical solution makes the combination of deep learning encoding tools with traditional video encoding schemes more efficient. Experimental results show that our approach achieves 0.01%, 0.31%, and 0.25% coding gain for the Y, U, and V components under the random access (RA) configuration on average.