Segmentation and visualization of the heart region for 3-D simulation of coronary intervention procedures

In fighting the coronary heart diseases, percutaneous coronary intervention is proved to be a powerful and reliable clinical procedure in the modern catheterization labs all over the world. Due to its minimally invasive characteristics, the procedure must be performed in the image-guided way, which makes this important skill very difficult to learn. To make the learning more accessible, a computer-aided surgical simulator is planned to be implemented in our lab. For now, the prototyping model is completed and the further validation is being conducted. In implementing the virtual anatomic environment, we aim to provide the trainee an intuitive visual effect so that the models of the organs bear resemblance to their counterpart of the human's. Besides the blood vessels per se, the surrounding organs seen in the real surgery also need to be visualized during the simulation. The heart is undoubtedly the most critical one among them. The segmentation of the heart is a challenging task because of the noisy and indistinct boundaries of the heart in the images due to the natural heart beating during the image acquisition. In this paper, an approach based on the active contours method is developed to fulfill this job. The experimental results demonstrate the effectiveness of our approach.