A proactive project is proposed which aims at the development of Pattern Recognition (PR) techniques and at studying
the impact of the application of these techniques to a wide range of of practical problems with great industrial or
commercial interest. PR is a well established field in Science and Engineering for more then 20 years, with applications
in most productive sectors. The present project explicitly focus on this multiplicity of areas in which clear benefits can be
obtained through the use of common PR techniques. The consolidated experience of the teams which participate in this
project guarantees the correct use, adaptation and/or development of those techniques which are most
appropriate for each application. Main PR technologies to be developed include Feature Selection and
Transformation and both
Supervised and Un-supervised (“Clustering”) Classification techniques. In all cases, the project will mainly focus on
approaches which, while being relatively simple, are proving extremely useful in practice. In particular, all the distance-
based techniques are worth mentioning. These techniques are quite flexible and general, since they do not even require
a vector representation of the objects involved. With regard to applications, eight specific practical problems are
considered: handwritten text recognition (OCR), prospective analysis of massive data sets (“Data Mining”), quality
control of ceramics,
fruit classification, marble classification, quality control of sole shoes, music analysis and medic applications.