Quality Resilient Deep Neural Networks

TL;DR

This paper introduces a mixture of experts ensemble approach for deep neural networks that enhances robustness to various image quality distortions by training specialized experts and a gating network to adaptively weight their contributions.

Contribution

The paper proposes a novel mixture of experts ensemble method with a gating network for robust image classification across different distortions, including weight sharing techniques for efficiency.

Findings

Ensembles outperform fine-tuned networks on distorted data.

Gating network effectively predicts expert weights without distortion info.

Weight sharing reduces model complexity with maintained performance.

Abstract

We study deep neural networks for classification of images with quality distortions. We first show that networks fine-tuned on distorted data greatly outperform the original networks when tested on distorted data. However, fine-tuned networks perform poorly on quality distortions that they have not been trained for. We propose a mixture of experts ensemble method that is robust to different types of distortions. The "experts" in our model are trained on a particular type of distortion. The output of the model is a weighted sum of the expert models, where the weights are determined by a separate gating network. The gating network is trained to predict optimal weights for a particular distortion type and level. During testing, the network is blind to the distortion level and type, yet can still assign appropriate weights to the expert models. We additionally investigate weight sharing methods for the mixture model and show that improved performance can be achieved with a large reduction in the number of unique network parameters.