Left ventricle segmentation based on real time MR image projections

Real-time image-guided procedures (manual or robot-assisted) are emerging due to potential improvement in patient management and reduction in the overall cost. To extract and incorporate information on-the-fly from intraoperative images requires new ultra-fast processing techniques. In this context, the work presents a method to process real-time MR images using signal intensity projections and extract boundary points on the tissues surfaces. The method was applied for segmenting blood-pool inside left ventricle of beating heart and extracting boundary points representing the endocardium wall. The robustness of the method was verified by comparing the position of the boundary points with the manual ground-truth. The near real-time processing speed allowed the methods to be applicable for intraoperative generation of dynamic tissues surfaces for guiding interventions in beating heart.