In order to improve the treatment of ischemic heart disease, the early detection of myocardial ischemia is crucial. This can for example be provided by the assessment of the myocardial blush grade, which is obtained by means of angiographic sequences. Usually, ECG-triggering is provided to ensure the correct alignment of diastolic images. However, further technical post-processing is absolutely necessary, aiming at co-registration of myocardium and epicardial coronary vessels. A fully automatic algorithm was developed which is capable of compensating motions of the heart induced by displacements such as movement of the patient table and respiratory motion etc. It includes the following three steps: segmentation, rigid registration and a final elastic registration. The segmentation is used to detect in each image the coronaries, which are filled with the contrast agent. Segmentation makes it possible to treat pixels representing dye inside the vessels differently to others during registration, enabling optimised registration of the vessels and of the corresponding myocardium. Registration errors caused by the moving vessels are compensated with a higher priority than those in the background. In order to further minimize the effect of the spreading dye, the images are registered in adjacent pairs. To achieve the registration of an image to the first one, all pairwise transformations preceding the relevant image are combined.
[1]
C. Steger.
SUBPIXEL-PRECISE EXTRACTION OF LINES AND EDGES
,
2000
.
[2]
Helmut Kücherer,et al.
Quantification of Myocardial Perfusion for CAD Diagnosis
,
2007,
MedInfo.
[3]
Hartmut Dickhaus,et al.
A flexible registration and evaluation engine (f.r.e.e.)
,
2007,
Comput. Methods Programs Biomed..
[4]
C M Gibson,et al.
Relationship of TIMI myocardial perfusion grade to mortality after administration of thrombolytic drugs.
,
2000,
Circulation.
[5]
Jan Modersitzki,et al.
Numerical Methods for Image Registration
,
2004
.