Robust real time motion estimation for MR-thermometry

Magnetic resonance imaging (MRI) allows continuous temperature mapping inside the human body and is thus a promising tool to guide thermal therapies. For abdominal organs, a real time motion compensation is required to allow quantitative temperature measurements and to permit retro-active control of the heating device. Although optical flow based algorithms have shown promise for this role, intensity variations induced by the heating process may lead to misregistration. Here, temperature information is used to identify such areas and introduced as an additional constraint into the formulation of the optical flow problem of the original Horn & Schunck´s approach. When combined with fast MRI, the current temperature measurement can be used as a weighting factor applied to the intensity conservation condition of the subsequent motion estimation. The proposed method showed in simulation and on MR-images of an ablation experiment a better estimation accuracy than the original Horn & Schunck´s approach.