Subcellular Structure Segmentation Method in Microscope Image
With the deepening of medical research, more and more human health-related diseases will be solved through a series of fruitful results and medical applications in the life sciences. Subcellular structures, such as intracellular proteins, are widely used as biosensors to study genotypic phenotypic relationships and quantitative analysis of drug efficacy. Since the localization capability of green fluorescent protein is strong, and the number of subcellular structures is large and of different shapes, the green fluorescent protein in subcellular cells can be traced under the microscope to realize the segmentation method studied in this paper. Different image segmentation methods are studied, including edge-based image segmentation, watershed algorithm and K-means clustering algorithm. The expectation maximization algorithm is then modified to calculate the parameters in the proposed model. The processing of cell images can be performed by using mathematical morphology, the basic operations of which are also studied. Based on the parameter estimation theory, noise suppression can be achieved in the continuous cycle process, thereby achieving high quality of reconstructed images. The experimental results show that the proposed algorithm can accurately and quickly segment the fluorescent protein image, which fully meets the requirements of protein localization and tracking.