The Analysis System Of Cell Fluorescence Images Based On MATLAB GUI

Different cells have different morphological characteristics,and the information related to cell function can be obtained through fluorescence labeling.With the rapid development of microscopic imaging technology,the collection and analysis of cell fluorescence image become particularly important.In the process of analyzing and processing cell fluorescence images,most cell image analysis software rely on parameters such as cell area threshold and fluorescence brightness threshold to obtain cell location information,and the results obtained will have certain errors,which cannot accurately reflect the information contained in cell fluorescence images.Appropriate image segmentation technology can reduce human involvement and objectively realize automatic recognition and data analysis.However,due to the diversity and complexity of cell images,not all image segmentation techniques are used for the analysis of cell fluorescence images,and appropriate segmentation methods should be selected according to the characteristics of cells and the purpose of cell analysis.In this paper,two types of cell fluorescence image analysis methods are mainly described: the first type is the fluorescence image of nerve cells.It is necessary to analyze the number,area and length of synapses,and compare the fluorescence intensity in the same area under the green and red fluorescence channels.In order to solve the problem of the deviation of the analysis results caused by the noise generated by the imaging equipment during the operation,the linear filter,median filter and wiener filter are analyzed and compared,and it is found that the median filter has better filtering effect.The OTSU threshold segmentation method is used to process the median filtered image to obtain the fluorescence region contained in the image.Compared with the edge detection segmentation,the former is beneficial to the area calculation of the fluorescence region,and the results obtained by this method are closer to the results of manual calculation than the data analyzed by the Image J software,which proves the feasibility of this method.Through man-machine interaction,the location of synapses can be marked or regions can be selected by the frame,and the location of regions can be extracted by the point method of the plumb line,so that only part of the regions in the image can be analyzed.The second type is the fluorescence image of HEK cells,which calculates the proportion of the sum of the fluorescence brightness of the cell membrane and the sum of the fluorescence brightness of the cells and analyzes the over-expression of the proteins in the cells.Due to the phenomenon of cell adhesion or the instability in the process of staining and cell image acquisition,as well as the different shapes of HEK cells,it brings many difficulties for cell recognition and analysis.In this paper,the image segmentation steps are divided into adhesion cell segmentation and cell membrane segmentation.Aiming at the problem of cell adhesion,boundary tracking and watershed transformation have achieved certain results.Through analysis and comparison,it is found that watershed transformation is more suitable for HEK cell fluorescence images analysis.By combining the results of distance transformation and gradient image with the minimum coverage technique,the over-segmentation in the cell region can be reduced,and then the location of each cell in the image can be determined by watershed transformation.Then,based on the results of cell segmentation,the fluorescence region of the cell membrane was obtained,and the luminance values of the cell region and the fluorescence region were calculated.In order to enable the experimenter to analyze image related data more quickly,the MATLAB GUI was used to build the cell fluorescence image processing system,the GUI designed the visual interface of the system,and the Editor wrote the logical program of the image analysis process.The user could directly import the folder where the images to be processed were stored to facilitate the selection of images to be analyzed.Button related functions can be realized by clicking the button.The results of image analysis are directly displayed in the system panel and can be saved.This system can basically meet the demand of nerve cells and HEK cells fluorescence image analysis,and can deal with image characteristics similar to the two similar images or analysis demand,improve the efficiency of image analysis,can be automated,objective image analysis results,also for other cell image analysis process to provide certain reference.