Using Dynamic Mode Decomposition for Functional Lung Imaging

Fourier Decomposition and related techniques have demonstrated the viability of obtaining regional pulmonary functions. To this end, novel post-processing techniques have been previously proposed to obtain ventilation and perfusion related information from dynamic acquisitions. To improve upon these methods, here we propose the use of an advanced data processing framework based on dynamic mode decomposition (DMD) for functional lung MRI. Phantom and in vivo results indicate that DMD achieves similar performance compared to established techniques and improves robustness in cases with fewer number of measurements.