Study And Implementation Of The Algorithm For The Segmentation Of The Sub-cellular Structure In Microscopy Images

Nowadays, an increasing number of medical problems depend on the developmentof the biomedical. In the biomedical field, cell is treated as the basic unit for research.The direct observation and research on the sub-cellular structure in the cell is greatlyimproved by the development of biological dyeing techniques at the molecular level anddigital imaging devices. At the same time, they generate large amount of image data. Itis imminent to find fast, accurate and automated analysis methods for the image data.The segmentation and analysis of sub-cellular structure image labeled by GFP (greenfluorescent protein) is one of them.This thesis studies the segmentation and processing of sub-cellular structure imagelabeled by GFP. The article analyses the characteristics of the sub-cellular structureimage labeled by GFP, the preprocessing of the image, the image segmentationalgorithm and the algorithm assessment to the two-dimensional image of sub-cellularstructure, the image segmentation algorithm and3D visualization to thethere-dimensional image of sub-cellular structure. Specifically, the main content andinnovation of the article have:1. The analysis to the characteristics of the sub-cellular structure image labeled byGFPL: The thesis analyzes the characteristics of the sub-cellular structure image labeledby GFP, which includes the variability of the sub-cellular structure, the GFP fading, andthe blurred edges. These features make it different from the general cell imagesegmentation for sub-cellular structure image labeled by GFP.2. The image segmentation of sub-cellular structure image labeled by GFP: Thisthesis summarizes the common image segmentation algorithm, and then it presents anew image segmentation algorithm to the sub-cellular structure image labeled by GFP.It can be used to segregate the two-dimensional image of sub-cellular structure andthere-dimensional image of sub-cellular structure.3. The assessment of the image segmentation algorithm: The thesis summarizes theexisting evaluation method of medical image segmentation algorithm, presents anassessment method to the image segmentation algorithm presented in the thesis. 4. The3D visualization of the sub-cellular structure image labeled by GFP: Manyof medical images are three-dimensional slice images. The sub-cellular structure imageslabeled by GFP are the same images. The three-dimensional reconstruction can help usto study the distribution of proteins in the cell. The article describes the method of thesurface rendering, the volume rendering and the accurately drawn based on the voxel ofthree-dimensional structure, and realized the three-dimensional reconstruction of thesub-cellular structure image labeled by GFP.Finally, using the hybrid programming of the Matlab2011，Visualization Toolkit5.8and Visual Studio2008integrate the above algorithm. It implements thesegmentation and display the two-dimensional and there-dimensional image ofsub-cellular structure.