Compression Artifacts Removal Using Convolutional Neural Networks

TL;DR

This paper demonstrates that large, deep convolutional neural networks can effectively reduce JPEG compression artifacts, outperforming previous methods in reconstruction quality through innovative training techniques.

Contribution

The paper introduces a training approach for deep CNNs that significantly improves JPEG artifact removal, utilizing residual learning, skip architecture, and symmetric weight initialization.

Findings

Deep CNNs outperform previous methods in artifact removal.

Training with residual learning and skip connections is efficient.

Networks generalize well across different datasets and JPEG quality levels.

Abstract

This paper shows that it is possible to train large and deep convolutional neural networks (CNN) for JPEG compression artifacts reduction, and that such networks can provide significantly better reconstruction quality compared to previously used smaller networks as well as to any other state-of-the-art methods. We were able to train networks with 8 layers in a single step and in relatively short time by combining residual learning, skip architecture, and symmetric weight initialization. We provide further insights into convolution networks for JPEG artifact reduction by evaluating three different objectives, generalization with respect to training dataset size, and generalization with respect to JPEG quality level.