Research On Cell Image Segmentation Technology And Its Application Under Digital Holographic Imagin

Digital holographic microscopy is a new kind of unmarked microscopic imaging technology,containing object wavefront by recording hologram,and then through the computer for the holograms recorded numerical reconstruction can obtain the phase and amplitude of object to be tested(intensity)information,achieving treat three dimensional measurement of object information,thus has been widely used in study of cytology.The study and analysis of cell movement and morphology in cytology requires accurate detection of cells in cell images,which means cell segmentation.The cell phase image obtained under digital holographic microscope has the advantages of high contrast and easy recognition,so it is suitable for cell segmentation.The precision of cell segmentation will directly affect the subsequent research and analysis of cells.However,due to the complexity of cell images,the traditional image segmentation algorithm has not achieved good results in the direct application of cell images,and the vast majority of current cell segmentation algorithms are only suitable for some specific forms of cells.For some cells with special morphology,it can not achieve better results.Based on the above research background,the segmentation method of cell phase image obtained under digital holographic microscope and its derivative application are studied in this paper.Firstly,the theory of digital holographic imaging and the live cell digital holographic microscopic imaging system used in this paper are introduced,and the characteristics of cell phase images obtained by this system are analyzed.Then,the cell segmentation technology is studied and discussed,and some algorithms are compared experimentally.According to the experimental results and data,the advantages and disadvantages of each algorithm are analyzed and summarized.On the basis of the above research,an image segmentation method of adhesive cells based on improved watershed was proposed.The algorithm was divided into three steps:cell region extraction,single cell recognition and marked watershed transformation.Before applying EGT algorithm to watershed algorithm,the over-segmentation phenomenon of watershed algorithm could be suppressed.Experimental results show that this algorithm can achieve accurate segmentation of images of adherent cells.Finally,we studied the cell division some derivative applications and discusses and introduces the realization of the cell count method and the calculation method of cell characteristic parameters,and proposes a tracking algorithm based on cell division and detection,the algorithm can achieve the cell movement in the video track obtained under digital holographic microscope and record its trajectory.