Machine Learning Techniques for Computer Vision

Duration: 3.5 hours including breaks.

Summary:

The application of machine learning is one of the most active frontiers in computer vision. This tutorial will provide a graduate-level introduction to some of the key concepts of machine learning, with an emphasis on techniques having broad applicability in computer vision. It will assume a familiarity with probabilities, but will otherwise be largely self-contained.

Topics to be covered will include:

Probabilistic graphical models; inference and learning.
Mixture distributions; the EM algorithm; variational methods.
Continuous latent variables; generative models of images.
Discriminative techniques for classification; neural networks; support vector machines; boosting; object recognition.
Example applications and case studies in computer vision.

On the instructor:

Chris Bishop is Assistant Director at Microsoft Research Cambridge and is responsible for machine learning, computer vision and information retrieval. He also holds a Chair of Computer Science at the University of Edinburgh where he is a member of the Institute for Adaptive and Neural Computation in the Division of Informatics, and is a Fellow of Darwin College, Cambridge. His research is concerned with the development and application of probabilistic methods for inference and learning.

He is author of the widely used text "Neural Networks for Pattern Recognition" (Oxford University Press, 1995) and is currently writing a new text book on pattern recognition and machine learning, to be published by Springer at the beginning of 2005.