# EyeInvaS: Lowering Barriers to Public Participation in Invasive Alien Species Monitoring Through Deep Learning

## TL;DR

EyeInvaS is a deep learning system that allows citizens to monitor invasive species using mobile phone photos, improving public participation and accuracy in ecological protection.

## Contribution

A novel AI-powered system for public participation in invasive species monitoring, validated through real-world deployment and benchmarked deep learning models.

## Key findings

- EfficientNetV2 achieved 83.66% and 93.32% F1-scores on original and hybrid datasets, respectively.
- Recognition accuracy was highest when targets occupied 60% of the frame against simple backgrounds.
- EyeInvaS enabled mapping of Solidago canadensis in Huai’an, China, showing strong associations with riverbanks and roads.

## Abstract

Invasive species pose serious threats to global biodiversity, agriculture, and ecosystems. Public participation offers an effective way to achieve large-scale and long-term monitoring, yet limited professional knowledge often reduces identification accuracy. This study introduces EyeInvaS, an intelligent image recognition system that enables citizens to identify and monitor invasive species simply by taking photos with their mobile phones. Using over ten thousand images—collected from online sources and synthetically generated under different scales and backgrounds—we built nine representative recognition models based on transfer learning and identified the optimal model and target scale through comparative analysis. The integrated EyeInvaS system supports key functions such as field reporting, rapid recognition, geographic tagging, and data sharing. Its reliability was validated through real-world field investigations of Solidago canadensis in Huai’an, China. This study demonstrates how deep learning technology can empower public participation in ecological protection and improve the efficiency of early detection and monitoring of invasive species.

Invasive alien species (IASs) pose escalating threats to global ecosystems, biodiversity, and human well-being. Public participation in IAS monitoring is often limited by taxonomic expertise gaps. To address this, we established a multi-taxa image dataset covering 54 key IAS in China, benchmarked nine deep learning models, and quantified impacts of varying scenarios and target scales. EfficientNetV2 achieved superior accuracy, with F1-scores of 83.66% (original dataset) and 93.32% (hybrid dataset). Recognition accuracy peaked when targets occupied 60% of the frame against simple backgrounds. Leveraging these findings, we developed EyeInvaS, an AI-powered system integrating image acquisition, recognition, geotagging, and data sharing to democratize IAS surveillance. Crucially, in a large-scale public deployment in Huai’an, China, 1683 user submissions via EyeInvaS enabled mapping of Solidago canadensis, revealing strong associations with riverbanks and roads. Our results validate the feasibility of deep learning in empowering citizens in IAS surveillance and biodiversity governance.