Maximilian Dostmann

Development a script to extract both quantitative and qualitative data from over 700 EU agricultural project reports, which served as the base for policy recommendations. My tool located and pulled quantitative parameters (like location, duration, and participants) based on unique webpage characteristics, structuring the output into a CSV file. It also collected, classified, and converted relevant text passages into an open document format for further qualitative analysis.

Planning a two-day workshop on automating counting tasks in biomedical imaging for neuroscientists. I taught machine learning and thresholding based methods for automated object segmentation to PhD students, professors and other researchers. They mainly required automated pipelines to recognize nerve cells from microscopy images and extract parameters from them.

Creation of a machine learning algorithm based on the convolutional neural network Stardist to recognize and analyze chloroplasts in 3D confocal microscopy images. It involved selection and processing of a training data set, integrating and adapting the machine learning algorithm into a workflow as well as image pre- and post-processing using ImageJ and OpenCV.