Hi I am currently reading:(adsbygoogle = window.adsbygoogle || []).push({});

http://www.cs.cornell.edu/~asaxena/learningdepth/saxena_ijcv07_learningdepth.pdf

which deals with reconstructing depth from a single still image.

A gaussian multiscale markov random field is trained in a supervised context where the model is shown below:

http://img534.imageshack.us/img534/1259/combineda.jpg [Broken]

now this model is converted into a standard multivariate gaussian (indicated by the arrow) but how is that conversion possible? I have read that it basically is a matter of completing the square but is there some more detailed explanation for this somewhere besides:

http://en.wikipedia.org/wiki/Completing_the_square

which don't really describe the techniques used on markov random fields.

**Physics Forums | Science Articles, Homework Help, Discussion**

Join Physics Forums Today!

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

# Converting a Gaussian Markov random field

**Physics Forums | Science Articles, Homework Help, Discussion**