Image Quality Assessment Techniques Show Improved Training and Evaluation of Autoencoder Generative Adversarial Networks

TL;DR

This paper introduces a new multidimensional evaluation method for autoencoder GANs, using multiple image quality metrics to improve training and assessment, leading to better image generation performance.

Contribution

It proposes a novel evaluation framework combining three distance functions for autoencoder GANs, enhancing training and image quality assessment.

Findings

Models with new distance functions produce higher quality images.

Each distance function captures different image properties.

Evaluation criteria improve model assessment accuracy.

Abstract

We propose a training and evaluation approach for autoencoder Generative Adversarial Networks (GANs), specifically the Boundary Equilibrium Generative Adversarial Network (BEGAN), based on methods from the image quality assessment literature. Our approach explores a multidimensional evaluation criterion that utilizes three distance functions: an $l_1$ score, the Gradient Magnitude Similarity Mean (GMSM) score, and a chrominance score. We show that each of the different distance functions captures a slightly different set of properties in image space and, consequently, requires its own evaluation criterion to properly assess whether the relevant property has been adequately learned. We show that models using the new distance functions are able to produce better images than the original BEGAN model in predicted ways.