A New Method for the Detection of Siliceous Microfossils on Sediment Microscope Slides Using Convolutional Neural Networks

• Published in: Journal of geophysical research. Biogeosciences 2024-09, Vol.129 (9), p.n/a
• Main Authors: Godbillot, Camille, Marchant, Ross, Beaufort, Luc, Leblanc, Karine, Gally, Yves, Le, Thang D. Q., Chevalier, Cristele, Garidel‐Thoron, Thibault
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial intelligence; Artificial neural networks; Automation; Bacillariophyceae; Climate change; Computer Science; Computer Vision and Pattern Recognition; Counting methods; Data acquisition; Datasets; Deep learning; Diatoms; Earth Sciences; Environmental changes; Environmental impact; Fossils; Geographical distribution; Image acquisition; Image quality; Impact analysis; Light microscopy; Machine learning; Marine microorganisms; Microorganisms; Neural and Evolutionary Computing; Neural networks; object detection; Object recognition; Ocean basins; Optical data processing; Optical microscopy; Paleontology; Phytoplankton; Plankton; Rare species; Sample preparation; Sciences of the Universe; Sediment; Sediment samplers; Sediment samples; Sediment traps; Sediments; Spatial data; Temporal resolution; Training; Workflow
• ISSN: 2169-8953; 2169-8961
• DOI: 10.1029/2024JG008047

Diatom communities preserved in sediment samples are valuable indicators for understanding the past and present dynamics of phytoplankton communities, and their response to environmental changes. These studies are traditionally achieved by counting methods using optical microscopy, a time‐consuming process that requires taxonomic expertise. With the advent of automated image acquisition workflows, large image data sets can now be acquired, but require efficient preprocessing methods. Detecting diatom frustules on microscope images is a challenge due to their low relief, diverse shapes, and tendency to aggregate, which prevent the use of traditional thresholding techniques. Deep learning algorithms have the potential to resolve these challenges, more particularly for the task of object detection. Here we explore the use of a Faster Region‐based Convolutional Neural Network model to detect siliceous biominerals, including diatoms, in microscope images of a sediment trap series from the Mediterranean Sea. Our workflow demonstrates promising results, achieving a precision score of 0.72 and a recall score of 0.74 when applied to a test set of Mediterranean diatom images. Our model performance decreases when used to detect fragments of these microfossils; it also decreases when particles are aggregated or when images are out of focus. Microfossil detection remains high when the model is used on a microscope image set of sediments from a different oceanic basin, demonstrating its potential for application in a wide range of contemporary and paleoenvironmental studies. This automated method provides a valuable tool for analyzing complex samples, particularly for rare species under‐represented in training data sets. Plain Language Summary Microfossils preserved in ocean sediments are studied to explore the impact of climate change on planktonic communities. The usual way to count these microfossils is slow and requires an expert to identify them on microscope images. In this study, we explore how artificial intelligence can be used on microscope images to detect the microfossils produced by one particular group, diatoms. Our results show that models can be trained to identify these objects, including the ones that were not specifically shown to the model during the training phase. However, the quality of the microscope image, and of the sample preparation beforehand, can affect how well the model works. This new protocol has good potential to be used on diatom imag