ASL is a novel non⁃invasive modality in functional Magnetic Resonance Imaging ( MRI) . ASL is capable of directly reflecting the Cerebral Blood Flow (CBF) of scanned patients, and is helpful to doing two things:(1) determining whether the scanned patient suffers from the dementia disease or not; ( 2 ) if the patient does suffer from the disease, determining the severity of the disease. However, since the spectral resolution of ASL is not high and problems of signal cross⁃contamination as well as voxel heterogeneity exist in this modality, the problem of Par⁃tial Volume Effect ( PVE) is commonly seen in ASL images. Moreover, PVE can easily deteriorate ASL signals, making the measured CBF of patients inaccurate and badly influencing their disease diagnosis thereafter. In this pa⁃per, a novel method solely based on single voxel information is introduced to correct PVE in ASL. Extensive experi⁃ments are conducted to suggest that this new method is not only capable of handling existing problems of blurring and brain detail loss which are commonly generated in conventional PVE correction methods but is also helpful to improving the precision of dementia disease diagnosis, from the statistical point of view.%动脉自旋标记磁共振图像是一种新颖的无侵入式的功能性磁共振图像。这类图像可以直观测量患者大脑血流量，对揭示患者是否患有老年痴呆症及判断相应的病症程度十分有效。然而，动脉自旋标记图像本身的分辨率不高，再加上扫描过程中不可避免的信号相叉污染和像素异质性等问题，使得部分容积效应在该类图像中普遍存在。部分容积效应会造成动脉自旋标记图像中信号还原失真，进而影响患者大脑血流量测量，对其病症判断带来不利影响。在文章中，一种基于单像素点信息的新颖算法被提出，改善动脉自旋标记磁共振图像中的部分容积效应。大量的统计比较实验表明，该算法不仅能解决国际上现行的改善算法中不可避免的改善结果过度模糊、丢失大脑细节信息问题，还能对准确判断患者老年痴呆病症程度能起到积极作用。
展开▼
