Feature analysis of wrapped phases of nucleated cells in polar coordinates

Digital phase microscopy has made great progress in the study of morphology and dynamics of biological cells, especially the homogeneous cells. For the heterogeneous nucleated cells with relatively complex internal structures, there are still few methods of phase unwrapping suitable for observation and measurement. In this paper, a method is proposed to identify the substructure of cells only with their wrapped phases. According to the structural features of real nucleated cells, their similar models are established firstly, and then their wrapped phases are obtained in Cartesian coordinates and polar coordinates. Based on the relationship between the parameters of the two coordinate systems, the pattern recognition and image processing techniques are used to analyze the parcel phase of the single-core and multi-core, regular and irregular kernels, and the central and boundary kernel models. Then, the texture shape distribution, the texture overlap condition, the opening and closing state of the equal phase line, the peak number and the peak position are selected as reference features, and these reference features are combined and optimized as the judgment condition for identifying the nuclear structure model, and the wrapped phase in polar coordinates is screened and identified, which effectively distinguishes three typical types of nuclear cell models. It provides a new, unpacked method for the classification and identification the substructures of the nucleated cells.