Case Study: Ludwig Maximilian University Gene Center achieves faster, lower-cost cryo-EM sample screening with MathWorks

Ludwig Maximilian University Develops Deep Learning System to Measure Ice Thickness in Cryo-EM Samples

Ludwig Maximilian University Gene Center worked with MathWorks to tackle the slow, expensive process of screening cryo-EM sample quality, especially measuring ice thickness in frozen grids. Using MATLAB and related tools such as Deep Learning Toolbox and Image Processing Toolbox, the team aimed to automate sample assessment from color camera images instead of relying on overbooked, costly electron microscopes.

MathWorks helped the LMU team build ANNICAS, a standalone deep learning application that combines image segmentation and classification to detect grid squares and classify ice thickness quality. The system reduced screening time dramatically, allowing 20 samples to be screened in 5 minutes instead of a full day, while also cutting microscope operating costs and saving researcher time.