2821

View Abstract / View Presentation

Sub-second Accurate Myelin Water Fraction Reconstruction with UNET

Jeremy Kim^1,2, Thanh Nguyen², Jinwei Zhang², and Yi Wang²
¹Stanford University, New York, NY, United States, ²Weill Cornell, New York, NY, United States

We developed UNET neural network for sub-second, accurate, and reproducible extraction of myelin water fraction map from FAST-T2 multi-echo T2 decay data.

Figure 1: Representative case showing the UNET’s prediction against the ANN’s prediction

Figure 2: RMSE for UNET and ANN Predictions