Electrolysis of water into oxygen and hydrogen is a cornerstone of modern energy storage, electric mobility and the transition towards a net-zero-emissions industry. Maximizing the efficiency of this technology is key to its economically viable wide scale adoption. One approach to both reduce the costs and improve the overall efficiency is to enhance the bubble detachment from the electrode. As the growing bubbles occupy the electrode surface, the rate of the electrochemical reaction is reduced. Therefore, the development of electrodes with improved bubble detachment is highly desirable. Analyzing the size distribution of the detached bubbles enables our partners to infer essential characteristics of their electrode designs. To this end, they acquire high-speed images of rising bubbles. As part of our collaboration we develop AI-based instance segmentation methods and blur estimation algorithms to automatically and scalably extract bubble sizes.