g-factor guided denoising for SENSE reconstructed MR images with edge restoration
Vijayakumar S, Huang F, Akao J, Duensing G
Owing to the geometry factor (g-factor), there is a great amount of non-uniformity in the distribution of noise in SENSE reconstructed images. Edges and noise are both high frequency information, low pass filters for denoising cannot distinguish between edges and noise. In this work, we apply a g-factor guided total variation filtering technique to denoise the images and add back the removed information (which includes both noise and edge information) weighted by g-factor to strike a balance between the smoothing and the edge preservation.