Converting a Gaussian Markov random field

  • Context: Graduate 
  • Thread starter Thread starter mort.motes
  • Start date Start date
  • Tags Tags
    Field Gaussian Random
Click For Summary
SUMMARY

The discussion focuses on converting a Gaussian multiscale Markov random field into a standard multivariate Gaussian, as outlined in the paper "Learning Depth from a Single Still Image" by Saxena et al. The conversion process involves completing the square, but participants seek a more detailed explanation specific to Markov random fields. The discussion references additional resources, including a paper on Bayesian approaches to image generation using Markov random fields, which may provide further insights.

PREREQUISITES
  • Understanding of Gaussian multiscale Markov random fields
  • Familiarity with the concept of completing the square in mathematics
  • Knowledge of multivariate Gaussian distributions
  • Basic principles of image generation techniques
NEXT STEPS
  • Research the mathematical foundations of Gaussian multiscale Markov random fields
  • Study the process of completing the square in the context of statistical models
  • Explore Bayesian approaches to image segmentation and generation
  • Examine the implications of multivariate Gaussian distributions in machine learning
USEFUL FOR

Researchers, data scientists, and machine learning practitioners interested in image processing, particularly those focusing on depth reconstruction and Markov random fields.

mort.motes
Messages
1
Reaction score
0
Hi I am currently reading:

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

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.
 
Last edited by a moderator:
Physics news on Phys.org
I'm not really familiar with image generation using Markov random fields. The following paper discusses approaches using a Bayesian approach. Perhaps it will be useful to you.

http://www.scss.tcd.ie/JiWon.Yoon/papers/MRF/Image%20segmentation%20image.pdf
 
Last edited by a moderator:

Similar threads

Undergrad Implementation of Correlated Gaussian Random Fields Model
  • · Replies 1 ·
Replies
1
Views
1K
Graduate 2D space and 1D time evolution of a random field
  • · Replies 5 ·
Replies
5
Views
2K
Graduate Classical interacting random field models
  • · Replies 24 ·
Replies
24
Views
7K
Admissions Need advice for my personal statement
  • · Replies 27 ·
Replies
27
Views
5K
Graduate BRS: Random Comments on Some Recent PF Threads
  • · Replies 94 ·
4
Replies
94
Views
14K
Mathematica This Week's Finds in Mathematical Physics (Week 218)
  • · Replies 1 ·
Replies
1
Views
3K
Mathematica This Week's Finds in Mathematical Physics (Week 258)
  • · Replies 4 ·
Replies
4
Views
10K