Wang, X.; Cai, B.; Jia, Z.; Chen, Y.; Guo, S.; Liu, Z.; Wan, X.; Zhang, F.; Hu, B.

Mitochondria are crucial organelles within cells, playing key roles in various biological processes, particularly in energy conversion, cell death, and signal transduction. Mitochondria consist of an outer membrane and inner cristae, the latter being essential for energy conversion. Consequently, quantitative analysis of the inner cristae not only reveals the functional state of mitochondria but also highlights its role in cellular metabolism and pathological conditions. However, there is currently a shortage of effective tools. This paper introduces MitoStructSeg, a deep learning-based platform for the segmentation and quantitative analysis of mitochondrial structures. Among these, the AMM-Seg model is proposed for mitochondrial structure segmentation, surpassing current state-of-the-art (SOTA) methods. Quantitative analysis of segmentation results elucidates the relationship between mitochondrial health and cristae structure. In addition, a user-friendly open source tool is available.