A 3-D space–time motion detection for an invariant image registration approach in digital subtraction angiography

A serious problem encountered in digital subtraction angiography (DSA) is the presence of artifacts due to patient motion. The resulting artifacts which arise from the misalignment of successive images in the sequence frequently reduce the diagnostic value of the images and lead to misdiagnosis or rejection of a DSA image sequence. A possible solution is the use of image registration techniques. In this paper, an invariant approach to elastic registration based on local similarity detection according to a combined invariants-based similarity measure, in which the images are locally registered and elastically interpolated, is presented. To improve the registration, a 3-D space–time motion-detection technique for extracting movement point pairs in a sequence of X-ray images is proposed. It is shown that such a technique is very efficient to correct for both global and local motion artifacts. The proposed algorithm for this technique has been successfully applied to register several clinical data sets including coronary applications. The results show that a fast and accurate image registration is achieved and that the algorithm is robust, which makes it clinically applicable.