The present invention relates to a method of analyzing PET (positron emission tomography) for diagnosis of dementia, and more specifically, sagittal plane and coronal view from an axial plane image obtained from a PET image of a axial plane of the brain. After automatically generating a coronal plane image, the artificial intelligence learning results are used for these three cross-section images to automatically classify and analyze normal, early, middle, late or normal, dementia stages for dementia. It relates to a deep learning-based PET image analysis method for diagnosis and prediction of dementia. In the present invention, a plurality of axial plane images are generated by taking a single layer of the axial plane of the brain, and then image interpolation is performed on the plurality of axial plane images to generate a 3D model, and the 3D model Sampling the sagittal and coronal planes at regular intervals to generate a plurality of sagittal and multiple coronal images and performing learning using a machine learning method. A PET image analysis method comprising a reasoner performing a diagnosis of dementia performing a process of inferring a degree, Providing a plurality of predetermined axial plane images； Generating a predetermined sagittal image and a coronal image, respectively, based on the predetermined axial plane image using a 3D model generator； The predetermined axial plane image, the predetermined sagittal plane image, and the predetermined coronal plane image are provided to an axial plane inference machine, a sagittal plane inference machine, and a coronal plane inference machine, and the axial plane inference machine, the sagittal plane inference machine, and And performing a diagnosis of dementia by inputting the results of the coronal inference machine into the selector, Normal, initial, and dementia output values for the axial image , , Define each as Normal, initial, and dementia output values for the sagittal image , , Define each a
