Comparison of image registration methods for composing spectral retinal images

Spectral retinal images have significant potential for improving the early detection and visualization of subtle changes due to eye diseases and many systemic diseases. High resolution in both the spatial and the spectral domain can be achieved by capturing a set of narrow-band channel images from which the spectral images are composed. With imaging techniques where the eye movement between the acquisition of the images is unavoidable, image registration is required. As manual registration of the channel images is laborious and prone to error, a suitable automatic registration method is necessary. In this paper, the applicability of a set of image registration methods for the composition of spectral retinal images is studied. The registration methods are quantitatively compared using synthetic channel image data of an eye phantom and a semisynthetic set of retinal channel images generated by using known transformations. The experiments show that generalized dual-bootstrap iterative closest point method outperforms the other evaluated methods in registration accuracy, measured in pixel error, and the number of successful registrations.