General information

There are huge amounts of data, signals, images, documents, multimedia documents, and so on. There is a need to manage the information somehow, but the pieces of information are often vague or partly missing. The research problem is: how to detect the weak signals from the available data?

The most common words describing our research are knowledge mining, data mining, knowledge management and information retrieval. By other words we are extracting pieces of information from signals and from complicated multimedia documents. Our research is mainly financed by Tekes and company projects.

The core competence of the research group is signal processing and mining of data originated from diverse sensors. The research work of the group can be characterized as data-driven research, which focuses on developing effective methods and algorithms to solve current real-life problems. The main research topics vary from signal modeling, information fusion, and detection to clustering, classification, and pattern recognition. In the past few years, the research has centered on data mining, outlier analysis, and online time series analysis from large sensor signal databases. Various sensor systems and multimedia applications, such as radars, multisensor systems, and ubiquitous multimedia, have been studied.