The methodological research of the LUT team has focused on approximative assimilation methods for high dimensional systems, both using dimension reduction methods and developing low-memory algorithms.  The applications include, in addition to weather forecasting,  satellite observations of the environment, forests and water systems, as well as remote tracking of flying objects. A recent direction is to develop rigorous statistical methods to quantify the variability and parameters of chaotic dynamical systems.

Professor Samuli Siltanen leads a research group concentrating on computational inversion methods for medical imaging, industrial applications, and art. There are three core topics.

  1. Reconstruction methods for X-ray tomography with limited data. In such imaging we record X-ray images of a patient or object along different directions of view and design a mathematical computer program for recovering the internal structure. Limited-data problems arise in medical imaging when the radiation dose to the patient needs to be kept very low, and in inspecting weldings in power plants when there are geometric restrictions for the view directions.
  2. Electrical imaging based on probing a patient or object with harmless electric currents. The main application area is to find out if a stroke victim is suffering from bleeding in the brain (hemorrhage) or from a blood clot preventing blood flow to the brain (ischemic stroke). The symptoms are the same in both cases, but the right treatment for ischemic stroke is dangerous to hemorrhagic patients. The imaging task is highly nonlinear and unstable, calling for special mathematical inversion techniques.
  3. Developing image processing methods that digital artists find useful for their work.