Implicit Neural Video Compression

TL;DR

This paper introduces implicit neural representations for video compression, representing each frame as a neural network and using learned quantization to reduce bitrate, simplifying neural video codecs and demonstrating feasibility.

Contribution

It presents implicit pixel flow (IPF), a novel neural video compression method that avoids pretrained networks and complex warping, enabling efficient full-resolution video compression.

Findings

Effective compression of full-resolution videos using neural implicit representations.

Avoids need for pretrained models and complex warping operations.

Demonstrates feasibility of neural implicit video compression.

Abstract

We propose a method to compress full-resolution video sequences with implicit neural representations. Each frame is represented as a neural network that maps coordinate positions to pixel values. We use a separate implicit network to modulate the coordinate inputs, which enables efficient motion compensation between frames. Together with a small residual network, this allows us to efficiently compress P-frames relative to the previous frame. We further lower the bitrate by storing the network weights with learned integer quantization. Our method, which we call implicit pixel flow (IPF), offers several simplifications over established neural video codecs: it does not require the receiver to have access to a pretrained neural network, does not use expensive interpolation-based warping operations, and does not require a separate training dataset. We demonstrate the feasibility of neural implicit compression on image and video data.