The invention discloses a progressive type mild
cognitive impairment identification method based on
neuroimaging, and belongs to the technical field of
computer image processing. The MRI (
Magnetic Resonance Imaging) graph and the PET (
Positron Emission
Tomography) graph of a
test sample are downloaded from an ADNI (Alzheimer's
Disease Neuroimaging Initiative)
database and are subjected to preprocessing and sample screening to obtain N groups of sample images; the AAL (Anatomical Automatic Labeling) template of the human is selected to independently manufacture 90 cerebral region templates for the sample images, and the
grey matter voxel value of a corresponding cerebral region is obtained to obtain N*180-dimensional data; and finally, a second level integration classifier is constructed,
feature dimension reduction is carried out on the obtained data, a reduced dimension is subjected to optimization, and the data is applied to the second level integration classifier to carry out classification identification on progressive type MCI (Mild
Cognitive Impairment) patients and non-progressive type MCI patients. The data is subjected to the dimension reduction
processing by a
random projection method, then, the data is applied to the second level integration classifier, classification accuracy is 74.22%, sensitivity is 66.25%, specificity is 82.19%, operation speed is improved, and the classification accuracy is improved.