Quantitative Image Analysis System For Mitochondria Dynamics

Mitochondria dynamics are a focus of cell biologists. The dynamic process relates closely to physiological cell conditions and human disease. The extraction of mitochondrial structural features and quantitative analysis of mitochondria morphology is a key parameter to access mitochondria dynamic process under different cell conditions. Nowadays morphology is usually analyzed by visual inspection, manual tracing and interactive software processing. These methods are tedium and slowness, and cannot effectively differentiate different shape. The present work was motivated by the need to achieve fast, objective, visible mitochondria quantitative analysis system. Fully automatic quantitative analysis system with a few parameter inputs is developed.First, proper morphology feature descriptors of mitochondria structure are proposed. Then, tubular structural image segmentation algorithms are reviewed. The mitochondrial quantitative analysis system is based on seed point selection, multi-directional template response on 16 directions, tracing algorithm and mitochondrial feature extraction. Its results include serial number labeled centerline of mitochondrial structure and mitochondrial morphological features.Mitochondrial structural features from fluorescence images of smooth muscle cells isolated from male SD rats was analyzed using the proposed algorithm. The new algorithm is more accurate and more efficient than Interactive method. The key feature descriptors include length, average wide, area, the location of mitochondria network node point and end point and so on. On the other hand, the new seed points selection method, new tracing direction detection method and the matrix storing feature information are proposed based on the existing tracing method. The new algorithm provides a powerful quantitative tool for mitochondrial dynamics research and can easily be extended for additional analysis. It can also serve as a useful tool for feature extraction of similar tubular and reticulum structures.