Research On Suppression And Compensation Of Respiratory Motion In Cardiac Imaging

Imaging technology has been widely used in the diagnosis, treatment planning and image guidance intervention of cardiac diseases, especially cardiac vascular diseases. Clinically acquired interventional cardiac images are always polluted by motion artifacts caused by cardiac and respiratory motion. The diagnostic accuracy of cardiac diseases and interventional therapy effect may be reduced. Cardiac motion artifacts can be suppressed by the ECG(electrocardiogram) gating. Respiratory motion is complicated which refers to the expansion and contraction of the thorax to breathe in and out. It may cause the heart to translate in 3-D. Thus, respiratory motion seriously affects cardiac image acquisition and image-guided intervention.Based on a comprehensive analysis of the generation and expression of respiratory motion artifacts in cardiac image sequences, an off-line gating method and direct compensation method are proposed to suppress respiratory motion artifacts in free-breathing interventional cardiac image sequences. Gating method is based on manifold learning to suppress respiratory motion artifacts. First, the dimensionality of the data matrix representing the image sequence was reduced with manifold learning algorithm. Then, gating frames at the same respiratory phase were detected. Direct compensation method compensates respiratory motion component through the establishment of cardiac or coronary vessel wall rigid motion model during respiratory cycle. No frames are discarded and the integrity of the image sequence is ensured. The experimental results with intracoronary ultrasound and intracoronary optical coherence tomographic images show that the respiration motion artifacts are effectively suppressed with our method. Additionally, this dissertation studies on respiratory motion model and lays the foundation for the respiratory motion artifacts' suppression by respiratory motion model.