A Group Variational Transformation Neural Network for Fractional Interpolation of Video Coding

TL;DR

This paper introduces a novel neural network-based fractional interpolation method for video coding that outperforms fixed filters, leading to significant bit savings and improved efficiency in motion compensation.

Contribution

The paper proposes a group variational transformation convolutional neural network (GVTCNN) that enhances fractional interpolation in video coding, offering a more adaptable and efficient approach.

Findings

Achieves 1.9% average bit saving over HEVC

Up to 5.6% bit saving in low-delay P mode

Improves fractional interpolation efficiency in motion compensation

Abstract

Motion compensation is an important technology in video coding to remove the temporal redundancy between coded video frames. In motion compensation, fractional interpolation is used to obtain more reference blocks at sub-pixel level. Existing video coding standards commonly use fixed interpolation filters for fractional interpolation, which are not efficient enough to handle diverse video signals well. In this paper, we design a group variational transformation convolutional neural network (GVTCNN) to improve the fractional interpolation performance of the luma component in motion compensation. GVTCNN infers samples at different sub-pixel positions from the input integer-position sample. It first extracts a shared feature map from the integer-position sample to infer various sub-pixel position samples. Then a group variational transformation technique is used to transform a group of copied shared feature maps to samples at different sub-pixel positions. Experimental results have identified the interpolation efficiency of our GVTCNN. Compared with the interpolation method of High Efficiency Video Coding, our method achieves 1.9% bit saving on average and up to 5.6% bit saving under low-delay P configuration.