Abstract
This paper presents a new blood vessel recognition algorithm that matches principle shapes and major point features between a pair of images by repeatedly using the iterated closest point (ICP) algorithm. In the initial stage, a set of primary blood vessels is extracted from the images. A transformation for global adjustment is found by matching primary blood vessels. In the second stage, a transformation for local adjustment is found by matching corresponding point features in the overlap area. Both matching procedures are based on the minimization of the sum of squares of differences (SSD) between two images and are performed by ICP algorithm. This approach guarantees both global and local alignment accuracy because global features are used to find the initial transformation and local features are used for optimal pixel alignment.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have