Efficient Morphometric Techniques in Alzheimer’s Disease Detection: Survey and Tools

Journal of computer sciences

N., Vinutha¹ 

关键词: Segmentation;    Registration;    Morphometry;    Classification;    Alzheimer?s Disease;   

DOI  :  10.3844/amjnsp.2016.19.44

学科分类：计算机科学（综合）

来源: Science Publications

【 摘 要 】

The development of advance techniques in the multiple fieldssuch as image processing, data mining and machine learning are requiredfor the early detection of Alzheimer’s Disease (AD) and to prevent theprogression of the disease to the later stages. The longitudinal and crosssectional images of elderly subjects were obtained from the standarddatasets like ADNI, OASIS, MIRIAD and ICBM. The subject imageobtained from the dataset, can be geometrically aligned to the templateimage through the process of registration. The registration techniques likeMutual Information Registration, Fluid registration, Rigid registration,Spatial Transformation algorithm for registration, Elastic Registration areselected based on type of transformation and similarity measures to suit therequired application. The registered images are then subjected to theprocess of segmentation in order to segment relevant tissues or desiredregion of interest that are significant in AD detection. The different types ofsegmentation techniques such as Tissue Segmentation, Atlas basedSegmentation, Hippocampus Segmentation and other segmentationtechniques have been discussed. The segmented images are then subjectedto morphometry techniques to identify the morphological changesdeveloped in an abnormal image. The different types of morphometrytechniques used are Voxel Based Morphometry (VBM), DeformationBased Morphometry (DBM), Shape Based Morphometry (SBM) andFeature Based Morphometry (FBM). But in recent years, the main focusof researchers is towards the FBM and SBM to overcome thedisadvantage of group analysis that existed in VBM and DBM. Furtherthe data is classified into healthy normal and AD by supervised,unsupervised or probabilistic methods.

【 授权许可】

CC BY