Mitochondria play a central role in cellular bioenergetics and in the regulation of apoptotic cell death. Mitochondrial morphology (shape and cristae architecture) is crucial to the understanding of apoptosis mechanisms and the subsequent development of therapies targeting age- and cancer-related diseases[1],[2]. There is a high demand in automated segmentation tools which can provide an objective quantitative information in a reasonable time frame[2]. The state of the art however is still dominated by manual segmentation tools[3]. Early attempts to address the challenges shown above are based on the machine learning framework[4].