StonERMax

Results 3 issues of StonERMax

The training's input is the clean image and noise. The training is along the forward direction , "Sdn->A*4->Gain->A*4", as the figure 3 in paper while all layers use the inverse...

Can the author share the pretrained network file ?