Is a speed-up possible

[Hunger-Prevails]

Hey Guys,

Huge thanks for your amazing tool to inpaint the depth maps. Currently I'm trying to use it to enhance a large amount of depth images, so that I can do 3D human pose estimation on that dataset. But I find your LRTV algorithm rather slow as it would basically run forever for an infinite number iterations just for a single image....
If I don't need too good a result and just want to preprocess the depth images as soon as possible, if there anyway to speed-up the process?