Automated Microfossil Identification and Segmentation Using a Deep Learning Approach

Ramos de Carvalho, L. E.; Fauth, G.; Baecker Fauth, S.; Krahl, G.; Camargo Moreira, A.; Peres Fernandes, C.; von Wangenheim, A.

doi:10.1101/661694

bioRxiv · 10.1101/661694

Automated Microfossil Identification and Segmentation Using a Deep Learning Approach

Ramos de Carvalho, L. E.·Fauth, G.·Baecker Fauth, S.·Krahl, G.·Camargo Moreira, A.·Peres Fernandes, C.·von Wangenheim, A.

View original source ↗Export metadata ↓

Abstract

The applicability of computational analysis to paleontological images ranges from the study of the animals, plants and evolution of microorganisms to the simulation of the habitat of living beings of a given epoch. It also can be applied in several niches, such as oil exploration, where there are several factors to be analyzed in order to minimize the expenses related to the oil extraction process. One factor is the characterization of the environment to be explored. This analysis can occur in several ways: use of probes, extraction of samples for petrophysical components evaluation, the correlation with logs of other drilling wells and so on. In the samples extraction part the Computed Tomography (CT) is of importance because it preserves the sample and makes it available for several analyzes. Based on 3D images generated by CT, several analyzes and simulations can be performed and processes, currently performed manually and exhaustively, can be automated. In this work we propose and validate a method for fully automated microfossil identification and extraction. A pipeline is proposed that begins in the scanning process and ends in an identification process. For the identification a Deep Learning approach was developed, which resulted in a high rate of correct microfossil identification (98% of Intersection Over Union). The validation was performed both through an automated quantitative analysis based upon ground truths generated by specialists in the micropaleontology field and visual inspection by these specialists. We also present the first fully annotated MicroCT-acquired publicly available microfossils dataset.View Full Text

Source connections

Published article ↗

Explore related subjects

paleontology Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Ramos de Carvalho, L. E., Fauth, G., Baecker Fauth, S., Krahl, G., Camargo Moreira, A., Peres Fernandes, C., von Wangenheim, A.. 2019-06-06. Automated Microfossil Identification and Segmentation Using a Deep Learning Approach. https://doi.org/10.1101/661694

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Durophagous adaptations in a new crocodyliform from the Santonian of Hungary and a systematic reassessment of European Cretaceous atoposaurids

Atoposauridae comprises a group of diminutive, semiterrestrial, brevirostrine neosuchian crocodyliforms that represents a characteristic component of European Jurassic-Cretaceous tetrapod faunas. Although frequently evaluated within the context of crown-group Crocodylia origins, their highly contested phylogenetic relationships continue to obscure their precise role in crocodylomorph evolutionary history. Particular ambiguity surrounds the monophyly of the genus Theriosuchus - with some hypotheses inferring long-term persistence of the lineage in Mesozoic Europe, whereas others interpreting select Cretaceous species as representatives of separate, even non-atoposaurid clades (e.g. Sabresuchus spp). We describe a new, stratigraphically intermediate taxon from the Santonian of Hungary based on isolated cranial elements, consistently placed as the sister taxon of Sabresuchus (Theriosuchus) sympiestodon from the Maastrichtian of Romania by our comprehensive set of phylogenetic analyses using three separate datasets. Sabresuchus tribodon sp. nov. is either recovered within Atoposauridae (global datasets) or Paralligatoridae (ingroup analysis) and can be distinguished from all other representatives of these groups by a specialized morphology consisting of massive, thick-enameled bulbous posterior maxillary teeth ornamented with apicobasal ridges, an extremely short rostrum with a reduced number of maxillary and dentary teeth, and the significantly upturned dorsal margin of the posterior dentary and surangular. Comparable anatomical adaptations are present in ecomorphologically convergent extinct crocodyliforms whose inferred feeding ecology included durophagy. However, this complete suite of characters has not been previously documented in either atoposaurids or paralligatorids. Our systematic reassessment results in an up-to-date taxonomy of Cretaceous European atoposaurids and tentatively confirms the polyphyly of Theriosuchus, though more definitive conclusions require combining individual, differently focused datasets. We reinterpret Aprosuchus ghirai from the Maastrichtian of Romania as a junior synonym of Sabresuchus sympiestodon and the putative European notosuchian Ogresuchus furatus from the Maastrichtian of Spain as an indeterminate atoposaurid.

paleontology↗

bioRxiv2026

Causal analysis of rock-biodiversity relationships over the global Phanerozoic challenges rock record bias and common cause hypothesis

Sampled marine biodiversity covaries strongly with sedimentary rock quantity over the Phanerozoic, but the respective contributions of different potential causes remain unsettled. Although regional studies show that rock-record bias accounts for more variation in sampled marine biodiversity than common-cause mechanisms, evaluating this relationship across the global Phanerozoic record remains difficult because fossil databases are inherently biased toward fossiliferous strata. To address this limitation, we compiled literature-derived measures of global marine biodiversity and sedimentary strata, along with an existing sea-level curve. Using an information-theoretic causal framework, we then evaluated the relative contributions of rock-record bias and common-cause hypotheses to Phanerozoic-scale variation in sampled marine biodiversity. Our results show a strong contemporaneous causality of marine strata on sampled biodiversity, indicating that rock record bias detected at regional scales extends to the global scale. However, neither rock record bias nor common-cause mechanisms accounts for most Phanerozoic-scale patterns in sampled biodiversity, and additional unmeasured factors likely contribute. Finally, we find tentative evidence that synergistic causality between geologically mediated sampling biases and forces shaping the underlying biological signal contributes to global Phanerozoic patterns in sampled biodiversity.

paleontology↗

bioRxiv2026

Increase in sea-level and shelf area positively influenced the development of Cenozoic biodiversity hotspots by modulating turnover rate in marine benthos

Marine biodiversity is unevenly distributed across the oceans, with exceptional concentrations of species richness confined to a few hotspots. Fossil and molecular evidence indicate that marine hotspots have shifted among ocean basins throughout the Cenozoic, but the mechanisms underlying their emergence and persistence have rarely been tested across multiple marine groups. Using standardized fossil occurrence data for molluscs, cnidarians, and foraminifera, we identified marine biodiversity hotspots through the past 66 million years. We then apply a multistep framework linking environmental drivers (sea surface temperature (SST), shelf area, and sea level) to clade-specific macroevolutionary processes (origination, extinction, and immigration) to evaluate the dynamics of these hotspots. Our results reveal contrasting evolutionary dynamics across marine biodiversity hotspots of different marine groups. Benthic groups: Mollusca and Cnidaria exhibit higher evolutionary turnover in hotspots than in non-hotspot regions, whereas planktic foraminifera maintain high diversity through reduced extinction and immigration rates. Increase in the sea level and shelf area positively contribute to the development of hotspot by increasing the origination rate in mollusca and cnidaria. Our study offers an integrative framework for understanding how marine biodiversity hotspots arise through a network of interacting processes: environmental drivers shape macroevolutionary dynamics, which in turn generate clade-specific diversity trajectories; together, these processes determine whether a particular region develops into a biodiversity hotspot.

paleontology↗