New Method of Classification to Detect Alzheimer Disease

This paper presents a new method of classification to detect Alzheimer Disease in any step: Early controls, Mild Cognitive Impairment (MCI) and Alzheimer Disease (AD). In this paper, we will present our method of classification, which is based on the results of the method of segmentation Level Set. The success of such a classification method is due to the segmentation method and the extraction of the area to be studied. Also, the descriptors used when extracting the form in order to give us better classification results. We will present our supervised classification method. The method consists in taking into account the 4 learning samples whose entry is closest to the new entry X, according to four distances: Euclidean, Manhattan, Hausdorff, AMED (Average Minimum Euclidean Distance). Base to estimate the output associated with a new input X. We tested our CAD with 75 subjects: 25 Normal (age ± SD=60 ± 8 years), 25 MCI (age ± SD=65 ± 8 years) and 25 Alzheimer (age ± SD=60 ± 8 years). The method proved an accuracy of 92% at Alzheimer Disease detection. Our method can be a useful tool to diagnose Alzheimer Diseases in any Step.