Motion tracking for clinical parameter derivation and adaptive flow acquisition in magnetic resonance imaging

A system for performing flow acquisition within a sequence of images being a sequence of magnetic resonance images, comprising:

• a medical imaging device for commencing an acquisition of a sequence of images;
• and
• an image processing device including a landmark detection module, a propagation module, and a structure fitting module, wherein
• the landmark detection module automatically detects one or more landmarks from within one or more long axis cardiac images of the sequence of images, the one or more landmarks including a valve point;
• the propagation module propagates the detected one or more landmarks across subsequent images of the sequence of images, including applying a deformable registration of the detected one or more landmarks across the subsequent images to predict a position of the one or more landmarks, wherein propagation comprises transforming a landmark from a starting phase to another phase using deformation fields and the deformation registration is computed using an inverse-consistent deformable registration algorithm;
• and
• the structure fitting module fits a geometric plane to the propagation of landmarks.

A method for determining tissue velocities within a sequence of images being a sequence of magnetic resonance images, comprising a processor:

• commencing an acquisition of a sequence of images;
• automatically detecting one or more landmarks from within one or more long axis cardiac images of the sequence of images;
• propagating the detected one or more landmarks across subsequent images of the sequence of images, including applying a deformable registration of the detected one or more landmarks across the subsequent images to predict a position of the one or more landmarks, wherein propagation comprises transforming a landmark from a starting phase to another phase using deformation fields and the deformation fields are computed using an inverse-consistent deformable registration algorithm;
• fitting a plane object to the propagation of landmarks in one or more long axis cardiac images;
• and
• determining one or more tissue velocities from the sequence of images within the fitted plane.