Simulation Evidence for use of a Denoising Auto-Encoder (DAE) to Improve Ultra-Low Field (64mT) MRI with a High Field (3T) Prior

In this work, the denoising autoencoder is applied to simulated low field, low resolution, low signal-to-noise images and used to recover the high field, high resolution, high signal-to-noise paired image. Different types of noise, gaussian and chi-squared, is added in simulation. We found that the denoising autoencoder worked slightly better for normally disturbed noise, but not in all cases. We found a linear trend between the model performance with RMSE and the standard deviation of the added noise. This work demonstrates the use of simple and robust denoising autoencoder to improve low field MRI.