| Summary: | 碩士 === 東吳大學 === 資訊科學系 === 96 === Positron Emission Tomography (PET) as an imaging modality in Alzheimer’s disease (AD) detection is attracting much of the attention because of its ability to provide an assessment of changes in brain metabolism. One of the key findings for patients with Alzheimer’s disease is characterized by the increased cerebrospinal fluid (CSF) due to the loss of gray matter tissues in the human brain. Therefore, CSF region size estimation, which determines the amount of CSF volume, becomes important for diagnosis of cognitive impaired AD patients. Despite of the growing interests in use of various types of biomarkers, the accurate estimation of CSF region is still hindered by the poor signal to noise ratio and partial volume effect of PET images in nature. This study aims to provide an automatic segmentation approach in determining the parameter settings for both sulcal and ventricular cerebrospinal fluid. The anatomically rich CT images are used to provide high-resolution complementary references. Experimental study using PET/CT images shows the promising results in future screening of AD patients with our proposed method.
|
|---|
