| Summary: | 碩士 === 中原大學 === 通訊工程碩士學位學程 === 107 === This paper presents an accurate method for the nuclear segmentation of three-dimensional conjugate focal microscope images. In the three-dimensional conjugate focal microscope image, since the nuclei are usually densely stacked or overlapped, the nuclei in the image are very close to each other, and individual nuclei cannot be identified to achieve visualization and quantitative analysis of various cell activities. Therefore, the nuclear segmentation method is a cell imaging study. One of the keys to success. Compared with the segmentation method of image processing, the current method of detecting the two contact nuclei in the image by using the protrusions and depressions on the contour of the nuclear nucleus has been widely discussed. These methods are accurate under certain conditions, but it is still impossible to detect two contact nuclei without depression in the core contour. Our method uses the size of the nuclei on the contours of the nuclei and the nuclei, diameters, etc. of the nuclei to detect the contact with the nuclei of the two-dimensional image. And we also use the nuclear size of the upper and lower 2D images and the nuclear eccentricity to achieve the correct six connections to ensure the reasonable size and shape of the resulting three-dimensional nuclear dimension. Taking the three-dimensional conjugate focal microscope image of mouse embryonic stem cells as an example, we detected the contact with the nucleus in the image to change the shape of the cells when the mouse embryonic stem cells were split.
|
|---|
