OLYMPUS BX53 motorized and fully equipped microscope with deep-learning technology (TruAI)

Universität Innsbruck

Innsbruck | Website

Open for Collaboration

Kurzbeschreibung

This is fully motorized Olympus (BX53) Upright microscope for the automated acquisition of multi-dimensional images, which enables the automatic scanning of several microscopic slides and stitching of picture sequences at once. The microscope is further linked to a computer which is equipped with image analysis software and deep-learning technology for automatic identification and measurement of the materials observed on microscopic slides. The microscope is equipped for transmission light, reflected light, polarization and fluorescence observations

Ansprechperson

Laurent Marquer

Research Services

This microscope and the related deep-learning technology provide automated scanning of thousands of pollen slides recorded manually during the last 40 years of pollen monitoring, upcoming new pollen slides for the pollen monitoring of the next years and decades, and also of thousands of slides from the pollen reference collection of the Department of Botany. These digital archives is used to measure the size of specific pollen types and study the variability of fluorescence characteristics of pollen from one plant species to another and through time. Slide scanning and automatic counting and identification by using the deep-learning technology make it possible to identify, count and calculate the concentration of pollen and other microscopic materials that are commonly observed on pollen slides, i.e. charcoal particles, black carbon content, fungal and algal spores and all types of microparticles.

The research infrastructure is "Open for Collaboration". Commercial collaborations are not possible.

Methoden & Expertise zur Forschungsinfrastruktur

This microscope and the related deep-learning technology provide automated scanning of thousands of pollen slides recorded manually during the last 40 years of pollen monitoring, upcoming new pollen slides for the pollen monitoring of the next years and decades, and also of thousands of slides from the pollen reference collection of the Department of Botany. These digital archives is used to measure the size of specific pollen types and study the variability of fluorescence characteristics of pollen from one plant species to another and through time. Slide scanning and automatic counting and identification by using the deep-learning technology make it possible to identify, count and calculate the concentration of pollen and other microscopic materials that are commonly observed on pollen slides, i.e. charcoal particles, black carbon content, fungal and algal spores and all types of microparticles. Over a long term, we will explore the possibility to develop an automatic identification and quantification of pollen grains and non-pollen palynomorphs.