• Home
• News
• Organized events
• Location

• Staff
• Students
• Technicians
• Visitors
• Former members

• Databases
• PhD Theses
• Projects
• Publications
• Software

• Access

Areas of interest

The research activity of the group concentrates mainly in the following areas of interest:

Pattern Recognition Algorithms

Pattern Recognition can be defined as the classification of input signals in different classes depending on their characteristics. To do this, it is necessary to separate those really significant properties from those that are irrelevant details for the classification. An example of this situation is the handwritten characters classification in the corresponding types of letters from images that are in general noisy (i.e., images that have stains or partial deletions). Other applications are the teledetection, speech recognition, fingerprints recognition, biomedical signal recognition, etc. Pattern Recognition has been treated from two different approaches:

• geometric methods, in which the signal is described in terms of a set of characteristics (number of curves, number of holes, etc)
• structural or syntactic methods, in which the signal is described in terms of the relationship between its components

Machine Learning

Machine Learning is an essential task for any recognition problem, since it permits to train the intelligent system from examples. The structuration of this knowledge as symbolic rules is the domain of grammatical inference, discipline in which the group has been specially active.

In this area of machine learning, it has been designed several algorithms that build correct models (finite state automata, stochastic automata, tree grammars, depending on the case) from examples of a language (which in each case they will be strings of symbols, trees, etc). Also, it has been developed methods that learn to translate from examples and permit also to use information about the characteristics of the language in order to accelerate the learning process.

Computer Music

Computer music is acquiring important relevance in the last few years. Automatic composition has been the area where traditionaly computers have played a significant role in the past, but now other areas related with human cognition have been introduced, like for example, music categorization, music performance, music indexing and retrieval, or perception of tone, rhythm, metrics, style, etc. In all these tasks pattern recognition and machine learning techniques can be of utility.

Our group tries to look for this kind of applications. Currently, projects on music style recognition, melodic similarities or automatic transcription of digital audio are being developed by our group.