A New Tracing Algorithm for Automatic Boundary Extraction from Coronary Cineangiograms

The need for reliable identification of vessel contours from X-ray image sequences within a limited computation time is still a challenge in medical image analysis. In the literature, only a few vessel boundary extraction methods are suitable to meet automatic and real-time constraints for capturing and processing coronary artery cineangiograms. Among many approaches, vessel tracing algorithms are known to be fast and robust for  practically detecting the vessel structures from live two- dimensional angiogram sequences. However, they often do not directly extract the boundary points and instead the locations of boundary points are achieved after identifying the correct position of center line points. This boundary detection scheme seems to be less efficient in speed demanding clinical applications. The team introduced a new algorithm for automatic tracing of vessel boundaries using an efficient estimation of local gradient vector. The results illustrate that the team method is a promising method for real-time vessel segmentation and linear feature extraction. Continue -->