AUTHOR=Zhang Wenjie , Bi Hongsheng , Wang Duansheng , Cheng Xuemin , Cai Zhonghua , Ying Kezhen 

TITLE=Automated zooplankton size measurement using deep learning: Overcoming the limitations of traditional methods

JOURNAL=Frontiers in Marine Science

VOLUME=Volume 11 - 2024

YEAR=2024

URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2024.1341191

DOI=10.3389/fmars.2024.1341191

ISSN=2296-7745

ABSTRACT=Zooplankton size is a crucial indicator in marine ecosystems, reflecting species diversity and the intricacies of ecological interactions. Traditional methods for measuring zooplankton size, involving direct sampling and microscopic analysis, are laborious and time-consuming.Moreover, the varying angles, orientations, and qualities of in-situ plankton imagery have posed significant challenges to early machine learning models. Addressing these challenges, our study introduces a novel, efficient, and precise deep learning-based method for zooplankton size measurement. This method leverages a deep residual network with an innovative adaptation: replacing the fully connected layer with a convolutional layer, thereby generating an accurate predictive heat map for size determination. We rigorously validated this automated approach against manual sizing using ImageJ, employing in-situ images from the PlanktonScope, and focusing on three zooplankton groups -copepods, appendicularians, and shrimps. An extensive analysis of 200 individuals from each group revealed that our automated method aligns closely with the manual process, showing a minimal average discrepancy of just 1.84%. This significant advancement presents a rapid and reliable tool for zooplankton size measurement, enhancing the capacity for immediate and informed ecosystem-based management decisions.