Skip to main content
. 2024 May 24;15:1392158. doi: 10.3389/fpsyt.2024.1392158

Figure 3.

Figure 3

Vision transformer analytical model. We prepared model A to E, which learned wavelet-transformed images, and each trained model was used to discriminate between patients with AD and HVs by dividing the dataset into five subsets (I to V). We used an ensemble model after predictions (Predicted IA~VE) into final prediction (Final Pred I~V), and combined final predictions.