Myocardial Motion Estimation Using Optical Flow with Multiple Constraint Equations

A new optical flow method is presented for estimating myocardial motion from MR image sequences. The concept of the local shape is introduced to generate four optical flow equations. Next, the interpolation operation is utilized to solve the four optical flow equations. Then a cost function is formulated from the intensity equation and gradient equations. The minimum method is considered to obtain an optimized position as the value of displacement. Finally, the optical flow method is tested with synthetic and experimental myocardial motion MRI data. There are two main contributions in this paper: the proposed estimation method can give a desired performance on both the eine and tagged MR image sequences, and for the simulation, we proposed synthesized eine and tagged sequences with the same ground truth motion field, which are used to evaluate the motion detection performances of the algorithms. Compared to classical methods, the results obtained improved displacement estimation and strain maps on eine and tagged images.