ICCV Course on Learning and Vision

October 12, 2003
Nice, France

 

Morning class
Learning and Vision: Generative Methods

Andrew Blake, Microsoft Research
Bill Freeman, MIT

Below are the topics covered, linked to the Powerpoint slides. See also the detailed course outline.

  • Maximum Likelihood Estimation (MLE) and Expectation Maximization (EM) (slides)
  • Principal Components Analysis (PCA), Factor Analysis (FA), and Transformed Components Analysis (TCA). (slides)
  • Markov Chains and Hidden Markov Models (HMM's) (slides)
  • Auto-Regressive (AR) dynamical models (slides)
  • Markov Random Fields (MRF's) (slides)
  • references.

 

Afternoon class
Learning and Vision: Discriminative Methods

Chris Bishop, Microsoft Research
Paul Viola, Microsoft Research