Logo image
在3D影像中以動態程式計算法連續分割之研究
Thesis

在3D影像中以動態程式計算法連續分割之研究

顏盟宜
Masters, National Tsing Hua University
2000

Abstract

自動分割連續分割對比度閾值法動態程式計算法 automatic segmentationsequential segmentationcontrastthresholdingdynamic programming
Abstract Our research focuses on the issue of automatic segmentation and sequential segmentation used in digital medical images. We use so-called dynamic programming to implement automatic segmentation by calculating the least total cost. First, we choose two points, the starting point and the terminal point, and set the total cost of terminal point as zero. Then to calculate the least total cost of the starting point and circle the desired region in the direction we calculate. Thus we have the desired segmented region. Second, we expand every point on the segmented boundary to form a donut-like region. The next image is read and the boundary is only searched using dynamic programming in the certain region. Co-operating with these parameters of the starting and terminal points we set up, we do have the second image segmented. Following these consecutive steps, the sequential segmentation is then achieved. From the results, the use of dynamic programming has excellent contrast of gray level between both sides of the boundary. The segmented shape is superior and suitable for sequential segmentation.Keywords: automatic segmentation, sequential segmentation, contrast, thresholding, dynamic programming

Metrics

1 Record Views

Details

Logo image