Research

Pattern Recognition is the scientific discipline which deals with theories and methodologies for designing machines capable to automatically recognise "objects" in noisy environments. Some typical applications are hand-written character recognition, remote-sensing image classification, people identification on the basis of biometrical features such as fingerprints.

Under the methodological viewpoint, our research interests in the field of Pattern Recognition are currenctly the following:

• Multiple classifier systems
• Reliability of pattern recognition systems (classification with the reject option)

We are also working on the following applications:

• Classification of remote-sensing images
• Advanced intrusion detection systems in computer networks
• Content-based retrieval from image databases
• Video surveillance systems
• Personal identity verification systems based on biometrics
• Document categorisation and Spam filtering