Compressed sensing using a deep adaptive perceptual generative adversarial network for MRI reconstruction from undersampled K-space data

Kun Wu, Yan Xia, Nishant Ravikumar, Alejandro F. Frangi

Abstract: Highlights•A novel deep perceptual feature guidance mechanism which can adaptively constrain the reconstruction by emphasising underlying and missed anatomical structures of multi-level.•De-aliasing robustness. Utilising multiple datasets and high undersampling rates for simulating the accelerated k-space data acquisition.•Subsequent analysis can benefit from using the reconstructed images, without significant influence on the biomarkers/anatomical measurements derived thereof, relative to those extracted from the original images.