Simultaneous Reconstruction of Respiratory and Cardiac Motion from Cine Magnetic Resonance Imaging

Abstract: Cardiovascular diseases are currently the leading cause of death in the world, which killed nearly 17 million people in 2011. For this reason, research in Cardiovascular diseases are of the up-most importance. In this thesis, real-time Cardiac Magnetic Resonance Imaging data is used to create simulated cardiac cycles for multiple phases of the respiratory cycle. By exploring image classification of both cardiac and respiratory cycles with a combination of cycle detection methods (Fast Fourier Transforms, Watershed Segmentation, K-means and locating maximas and minimas) and using a RANSAC method for robustness, interpolated volumes for each respiratory cycle can be created. The use of different interpolation methods are also explored to discover how to produce the best results. In conclusion a Monotonic Piecewise Cubic Spline Interpolation in combination with the use of an optimisation method, to select the most suitable images, proved to be the most accurate method to produce simulated cardiac cycles. The ejection fraction obtained at expiration, from the simulated cardiac cycle, has a value of 55.6962 +/- 1.6199% which is within the current standard normal range of 55-70\% determined at the same respiratory phase.