Pattern Recognition is a highly pervasive discipline, both for science and industry. It focuses on the creation of classifiers, that is, algorithms able to learn aspects of the reality that surrounds us and to make appropriate decisions when in the presence of new stimuli. Speech recognition, automotive, surveillance systems, quality control systems, recommender systems, search engines, social networks, interactive tools (Kinect, Wii) are just some of the many applications that rely on the presence of classifiers. The Pattern Recognition course is intended to provide the methodological principles at the basis of the classification, together with the most modern techniques that can solve problems until a few years ago unmanageable. In other words, the course aims to be the best compromise between theory and practice, making the student can solve problems with tangible and important techniques from solid theoretical point of view.
The course can be divided into two parts, the methodology and the application, which go hand in hand during the course.
- Recognition and classification
- Bayesian Decision Theory
- Parameters Estimation
- Nonparametric Methods of Parameters Estimation
- Linear and non-linear discriminant functions
- Extraction and feature selection, PCA, Fisher transform
- Expectation-Maximization Algorithm on mixtures of Gaussians
- Generative and discriminative methods
- Kernel Methods and Support Vector Machines
- Hidden Markov Models
- Methods for unsupervised classification (clustering)
- Pattern recognition for the analysis and recognition in images and videos
- Face recognition
- Video surveillance
- Richard O. Duda, Peter E. Hart, and David G. Stork. 2000. Pattern Classification (2nd Edition). Wiley-Interscience.
- Christopher M. Bishop. 2006. Pattern Recognition and Machine Learning (Information Science and Statistics). Springer-Verlag New York, Inc., Secaucus, NJ, USA.