Deep Learning Applied to SEM Images for Supporting Marine Coralline Algae Classification

The classification of coralline algae commonly relies on the morphology cells and reproductive structures, along with thallus organization, observed through Scanning Electron Microscopy (SEM). Nevertheless, species identification based often leads to uncertainty, due their general plasticity. Evolutionary environmental studies featured for ecological significance in both recent past Oceans need rely robust taxonomy. Research efforts towards new putative diagnostic tools have recently been focused cell wall ultrastructure. In this work, we explored a tool algae, using fine-tuning pretrained Convolutional Neural Networks (CNNs) SEM images paired morphological categories, including We considered four common Mediterranean species, classified at genus level (Lithothamnion corallioides, Mesophyllum philippii, Lithophyllum racemus, pseudoracemus). Our model produced promising results terms image accuracy given constraint limited dataset was tested two ambiguous samples referred as L. cf. racemus. Overall, explanatory analyses suggest high value calcification patterns, which significantly contributed class predictions. Thus, CNNs proved be valid support approach taxonomy algae.