Research on Registration of Human Eye Tissue Images and Its Application
Xiaoyang Huang, Boliang Wang, Ying Ju, Shaohui Huang, Jiezhen Xie, Liwei Huang · 2005
Some methods for registration of human eyes tissue images were described in this paper. Firstly, the algorithm of moment and principal axes was used to register the image sequence of eye's tissue slices in the rough. Then, a modified segmentation-counting registration algorithm was proposed to register the image sequence more accurately, which is not only as precision as maximization of mutual information registration algorithm but also more effectively than it. It was proved by experimentation that this algorithm can 10 times quicken the speed of the maximization of mutual information registration algorithm. Some kinds of optimization algorithms, such as Powell, Brent, Gloden Section and the Second-degree interpolation algorithms, had been applied to the registration of the image sequence. The results of the experimentations showed that the registration speed can be accelerated more than 300 times faster with these algorithms.