Deep motion estimation for parallel inter-frame prediction in video compression

TL;DR

This paper introduces a neural network-based method for learning binary motion codes that capture complex motion in video frames, enabling parallel decoding and outperforming traditional codecs at low bitrates.

Contribution

It presents a novel learned inter-frame prediction model with binary motion codes supporting complex motion and parallel decoding, improving compression efficiency.

Findings

Outperforms H.264 and H.265 codecs at low bitrates.

Supports parallel decoding instead of sequential flow estimation.

Achieves bit savings through 3D dynamic bit assignment.

Abstract

Standard video codecs rely on optical flow to guide inter-frame prediction: pixels from reference frames are moved via motion vectors to predict target video frames. We propose to learn binary motion codes that are encoded based on an input video sequence. These codes are not limited to 2D translations, but can capture complex motion (warping, rotation and occlusion). Our motion codes are learned as part of a single neural network which also learns to compress and decode them. This approach supports parallel video frame decoding instead of the sequential motion estimation and compensation of flow-based methods. We also introduce 3D dynamic bit assignment to adapt to object displacements caused by motion, yielding additional bit savings. By replacing the optical flow-based block-motion algorithms found in an existing video codec with our learned inter-frame prediction model, our approach outperforms the standard H.264 and H.265 video codecs across at low bitrates.