Microplastic Identification Using AI-Driven Image Segmentation and GAN-Generated Ecological Context

TL;DR

This paper introduces an AI-based approach combining image segmentation and GAN-generated data to improve microplastic detection in water samples, making analysis more accurate, cost-effective, and accessible.

Contribution

It presents a novel deep learning segmentation model trained with GAN-augmented data, significantly enhancing microplastic identification accuracy over previous methods.

Findings

Segmentation model achieved an F1-Score of 0.91 with GAN data.

Expert reader study showed 68% accuracy in distinguishing real from generated microplastics.

Model outperformed non-GAN trained models with an F1-Score of 0.82.

Abstract

Current methods for microplastic identification in water samples are costly and require expert analysis. Here, we propose a deep learning segmentation model to automatically identify microplastics in microscopic images. We labeled images of microplastic from the Moore Institute for Plastic Pollution Research and employ a Generative Adversarial Network (GAN) to supplement and generate diverse training data. To verify the validity of the generated data, we conducted a reader study where an expert was able to discern the generated microplastic from real microplastic at a rate of 68 percent. Our segmentation model trained on the combined data achieved an F1-Score of 0.91 on a diverse dataset, compared to the model without generated data's 0.82. With our findings we aim to enhance the ability of both experts and citizens to detect microplastic across diverse ecological contexts, thereby improving the cost and accessibility of microplastic analysis.