A Survey on Evaluation Metrics for Synthetic Material Micro-Structure Images from Generative Models

TL;DR

This paper reviews and evaluates various metrics for assessing the quality of synthetic micro-structure images generated by models, highlighting their limitations in the materials science context and encouraging improved evaluation methods.

Contribution

It provides an analysis of existing evaluation metrics on SEM images of micro-structures, identifying their shortcomings and proposing directions for better assessment tools.

Findings

Current metrics like Fréchet Inception Distance have limitations for micro-structure images.

Existing methods often fail to capture physically relevant features.

The study highlights the need for domain-specific evaluation metrics.

Abstract

The evaluation of synthetic micro-structure images is an emerging problem as machine learning and materials science research have evolved together. Typical state of the art methods in evaluating synthetic images from generative models have relied on the Fr\'echet Inception Distance. However, this and other similar methods, are limited in the materials domain due to both the unique features that characterize physically accurate micro-structures and limited dataset sizes. In this study we evaluate a variety of methods on scanning electron microscope (SEM) images of graphene-reinforced polyurethane foams. The primary objective of this paper is to report our findings with regards to the shortcomings of existing methods so as to encourage the machine learning community to consider enhancements in metrics for assessing quality of synthetic images in the material science domain.