An attempt to use Markov-Chain Monte Carlo methods to perform 3D phase
unwrapping modeled as a low-level vision factor graph (Markov Random
Field). Uses advanced MCMC techniques including parallel tempering, 
data-driven MCMC, and partial replica exchange. Does not perform
great, but serves as an interesting starting point. 

Implemented in: 
  core: mostly c++ 
  wrapper: python


see presentation/presentation.pdf for more