| Summary: | 碩士 === 國立嘉義大學 === 資訊工程學系研究所 === 97 === Confocal laser scanning microscope plays a big role in the visualization of 3-D cell. However, due to the nature and the acquiring environment of confocal laser scanning microscopes, the sampling results are prone to the influence of scattered light and spherical aberration. The resulting images are also, at times, unclear, which may lead to difficulty in extracting the contour of fluorescent cells. This study proposes a method base on active contour model to detect the cells. Duo to the limitations of active contour model, there are some problems especially low-contrast at cytoplast areas. This study proposes a series of image processing operations methods to overcome these problems. After segmentation, the proposed system reconstructs a smooth 3D cell by stacking a series of two-dimensional fluorescent images into volume data which can visualize protein activity, and ER by using computer graphics.
|
|---|
