A New Method for the Detection of Siliceous Microfossils on Sediment Microscope Slides Using Convolutional Neural Networks DOI Creative Commons
Camille Godbillot, Ross Marchant, Luc Beaufort

et al.

Journal of Geophysical Research Biogeosciences, Journal Year: 2024, Volume and Issue: 129(9)

Published: Sept. 1, 2024

Abstract Diatom communities preserved in sediment samples are valuable indicators for understanding the past and present dynamics of phytoplankton communities, their response to environmental changes. These studies traditionally achieved by counting methods using optical microscopy, a time‐consuming process that requires taxonomic expertise. With advent automated image acquisition workflows, large data sets can now be acquired, but require efficient preprocessing methods. Detecting diatom frustules on microscope images is challenge due low relief, diverse shapes, tendency aggregate, which prevent use traditional thresholding techniques. Deep learning algorithms have potential resolve these challenges, more particularly task object detection. Here we explore Faster Region‐based Convolutional Neural Network model detect siliceous biominerals, including diatoms, trap series from Mediterranean Sea. Our workflow demonstrates promising results, achieving precision score 0.72 recall 0.74 when applied test set images. performance decreases used fragments microfossils; it also particles aggregated or out focus. Microfossil detection remains high sediments different oceanic basin, demonstrating its application wide range contemporary paleoenvironmental studies. This method provides tool analyzing complex samples, rare species under‐represented training sets.

Language: Английский

Has Quaternary palynology reached its climax? DOI
Jan Barabach

The Holocene, Journal Year: 2025, Volume and Issue: unknown

Published: April 25, 2025

Throughout more than 100 years of its history palynology has become an independent discipline that is being applied in various fields from palaeoecology, aerobiology, forensic sciences to taxonomy. Bibliometric analysis allows distinguish different phases the evolution palynology. From initial phase, when first pioneer results were released, through building phase potential pollen been expanding into new research areas, mature which becomes a basic method with worldwide recognition many scientific fields. However, scientometric palynological documents suggests second decade 20th century, increasing trend number published stopped. This tendency observed most journals publishing content. may suggest reached climax. Moreover, last couple show slow but constant drop documents. During this period also decrease mean citation per document and year observed. poses question – does issue reflect only state or wider phenomenon touching other related as for example palaeoecology? On hand, bibliometric points out some positive aspects such increase international co-authorship co-authors indicates development specialization discipline.

Language: Английский

Citations

0
A New Method for the Detection of Siliceous Microfossils on Sediment Microscope Slides Using Convolutional Neural Networks DOI Creative Commons
Camille Godbillot, Ross Marchant, Luc Beaufort

et al.

Journal of Geophysical Research Biogeosciences, Journal Year: 2024, Volume and Issue: 129(9)

Published: Sept. 1, 2024

Abstract Diatom communities preserved in sediment samples are valuable indicators for understanding the past and present dynamics of phytoplankton communities, their response to environmental changes. These studies traditionally achieved by counting methods using optical microscopy, a time‐consuming process that requires taxonomic expertise. With advent automated image acquisition workflows, large data sets can now be acquired, but require efficient preprocessing methods. Detecting diatom frustules on microscope images is challenge due low relief, diverse shapes, tendency aggregate, which prevent use traditional thresholding techniques. Deep learning algorithms have potential resolve these challenges, more particularly task object detection. Here we explore Faster Region‐based Convolutional Neural Network model detect siliceous biominerals, including diatoms, trap series from Mediterranean Sea. Our workflow demonstrates promising results, achieving precision score 0.72 recall 0.74 when applied test set images. performance decreases used fragments microfossils; it also particles aggregated or out focus. Microfossil detection remains high sediments different oceanic basin, demonstrating its application wide range contemporary paleoenvironmental studies. This method provides tool analyzing complex samples, rare species under‐represented training sets.

Language: Английский

Citations

2