Figure 1
From: Mapping the association between tau-PET and Aβ-amyloid-PET using deep learning
Processing pipeline for a convolutional neural network (CNN)—a deep learning technique developed to process images—to determine the association between tau-PET topography and Aβ CL. Tau images and Aβ CL values form the input and output of the CNN, respectively. After training the CNN, a saliency map is generated for every subject and associated tau-Aβ CL mapping, in which every voxel is allocated a weight to indicate the importance of its tau uptake value in the estimation of the CNN estimated Aβ CL.
