With the capacity to generate large amounts of image data comes the need of ways to process and analyse the incoming data. To distinguish, identify and quantify microorganisms not only requires sufficient equipment (high magnification microscope) but foremost extensive taxonomic knowledge. This knowledge and skill is unfortunately in short supply and limited to a certain processing speed. If we are to process large amounts of data, for broad quantitative purpose, manual methodology becomes very expensive and time consuming.

Recent years advances in, and availability of, open-source libraries for machine learning (eg TensorFlow and OpenCV) has made development of object detection and classification models more accessible. Employing AI to assist in the identification and quantification of microscopic organisms would then enable for the processing of large amounts of images.

It is within the scope of the BysjöBot project to, in collaboration with taxonomy and AI experts, explore and evaluate the opportunities and limitations of AI-assisted identification and quantification of microorganisms. These tools will then be added to the workflow.

This webpage, with all its contents, is to be considered under contruction and not a final product.
Links and descriptions to related solutions and protocols will continously be updated and made available.