SYRACO (SYstème de Reconnaissance Automatique de COccolithes) is a software developed at CEREGE since 1995 and integrated into the MANTA platform, dedicated to the analysis, annotation, and automatic classification of microfossils through artificial intelligence. Initially, we used our own Convolutional Neural Network (CNN) developed at CEREGE and published in 1997 and 2004, which was used until 2017. Afterward, we adopted deep learning techniques for enhanced performance.

SYRACO allows researchers to efficiently manage image sets by providing tools for importing, collaborative annotation, training classification models, and advanced morphometric analysis.

Thanks to its deep learning tools (YOLO, ResNet, U-Nets), SYRACO facilitates the taxonomic identification of nannofossils and other microfossils, with interactive visualization features and data export capabilities tailored to the needs of the scientific community.

It is associated with the CoccoScan software, which controls the 6 automated microscopes of the MANTA Platform.

As an open-source project, SYRACO aims, in the near future, to unite researchers around a common database, thereby optimizing the use of paleontological data to better understand the evolution of the oceans and climate.