碩士 / 國立高雄第一科技大學 / 電腦與通訊工程研究所 / 99 / Alzheimer''s Disease is an age-related, non-reversible brain disorder that develops over a period of years. People who have lost brain cells in the hippocampus area of the brain are more likely to develop Alzheimer''s Disease. The traditional method needs an experienced physician to observe the hippocampal atrophy, which is inefficient and will lead to misjudgment and waste much time to hippocampal volume calculation when the physician lacks enough clinical experience. To this end, this study propose a automatic system of hippocampal volume measures using a wavelet-based dynamic boundary extension algorithm. The proposed system includes the hippocampus boundary detection and volume calculation. First, the hippocampus boundary detection is used to find the Area of Interest (AOI) of hippocampus automatically, which can reduce the processing time and avoid the unwanted interference. Next, we use Histogram Equalization to enhance the image of AOI. and wavelet transform is used to estimate the initial boundary of hippocampus. By manual locating the initial seed point and the dynamic boundary extension algorithm, we reconstruct the edge of hippocampus to calculate the total pixels inside the contour as the hippocampus area. To obtain the hippocampal volume of patients, the volume calculation is done by accumulating each image within the same group in the hippocampus area.
The dataset of this paper includes 10 patients. Converting by the three-dimensional reconstruction, each patient has 512 sagittal plane images. The experimental analysis is conducted by using the 331 images, which are selected from the 5120 images. Experimental results show that the successful identification of the hippocampus is up to 84%, while proposed system have the advantages of fast hippocampal volume calculation and easy to find hippocampus area even for people lack of experience.
| Identifer | oai:union.ndltd.org:TW/099NKIT5650043 |
| Date | January 2011 |
| Creators | Jyun-Jhang Chen, 陳俊璋 |
| Contributors | none, none, 汪桓生, 洪金車 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 84 |
Page generated in 0.014 seconds