Pattern Recognition
Pattern recognition in video sequences is a fundamental task to understand what is happening. Machine learning techniques are investigated to provide solutions to different video surveillance problems from middle level techniques (e.g. object recognition/classification) to high level techinque for the behaviour analysis (e.g trajectory analysis, anomaly detection, etc.)
Active Vision
Active Vision is the estabilished area of computer vision. Such an area aims to the interpretation of the scene by analysing images acquire by camera mounted on moving platfrom. Thus, active vision is related to the interaction between the oberver and the sensor to actively decide what,when and how to see. Hence, this field of research not only copes with the processing of the images but also with the control of the motion of the sensors and their configuration.
Resource Aware Networks
In context of sensor networks three main resources present strong limitations: a) Communication, b) Computational Power and c) Power Consumption. All these limitations become more evident with the increase of the number of sensors deployed by the network. These limitations require a sort of awareness of the system that can take measures to adapt its performance or operability in order to adhere to the constraints. In this field Dr. Micheloni is studyning techinques to adapt the network topology by switching on/off cameras and/or modifying their field of view.