| SSIP 2011 Home |
 |
|---|---|
Registration |
Vladimir Kolmogorov: MAP-MRF inference techniques in computer vision |
| Many problems in Computer Vision are formulated in the form of a random
field over discrete variables, such as Markov or Conditional Random
Field (MRF/CRF). Examples range from low-level vision such as image
segmentation, optical flow and stereo reconstruction, to high-level
vision such as object recognition. The goal is typically to infer the
most probable values of the random variables, known as Maximum a
Posteriori (MAP) estimation.
In this tutorial I will cover several algorithms for MAP-MRF estimation. I will focus on techniques based on the dual decomposition approach. In particular, I will consider tree-reweighted message passing and its precursor, max-product belief propagation. I will also discuss an alternative subgradient ascent approach, and illustrate it on some concrete computer vision problems such as computing correspondences between sparse image features. 
|
|
|
Page
last modified: |
Please send
comments to ssip2011@inf.u-szeged.hu |