Publication Detail Page

S. Lee. A study of ophthalmologic imaging distortion for retinal image registration and validation. PhD thesis, The University of Iowa, Iowa City, IA, 2008.

Abstract: The human retina is the only part of the central nervous system that can be imaged directly. Fundus camera imaging of the retina is widely used to diagnose and manage oph- thalmologic disorders. Optical imaging of the fundus typically has a limited field of view because of the narrow light entrance through the pupil and the spherical shape of eyeball. Thus in order to generate a larger view and potentially improve diagnostic accuracy, the images from separate regions of the retina can be combined. In this study, the optical fundus images are stitched to generate a single view of retina using image registration technique. Specifically, we assume the retinal image undergoes geometric coordinate distortion during the imaging process, which is due to the non-linear optics and the curvedness of retinal surface. Thus, a mathematical model of ophthalmologic imaging distortion is developed and applied to the retinal image registration. The experimental results using a large number of real fundus images show that the proposed distortion correction process can register retinal images with high accuracy. Especially for the mosaicing purpose, it becomes possible to register images in a cascade so the montage image is free of geometric distortion propagation. The proposed registration method is novel in that the registration is performed in the distortion-free coordinate system. Registering images in the distortion corrected coordinate system allows many advantages over the conventional registration methods. It can dramatically reduce the degree of freedom for image transformation so the method is well-suited to the partial overlap registration problem, i.e. combining retinal images. Once the distortion is recovered, the registration becomes a matching problem by linear operations. Eventually, the resulting montage by proposed method shows no coordinate warping. This study presents a physical coordinate distortion model of opthalmologic imaging so a set of simulated retinal images can be generated by user-controllable parameters. The simulated images can directly be utilized to validate the accuracy of registration algorithms. The validation framework proposed in this study enables to compare the registration accuracy even for the non-overlapping region.

Citation: BibTeX format Endnote format

Keywords: registration retina

Other publications by: S. Lee

Related journal papers:

B. Li, G. E. Christensen, G. McLennan, E. A. Hoffman, and J. M. Reinhardt. Establishing a normative atlas of the human lung: Inter-subject warping and registration of volumetric CT. Acad. Radiol., vol. 10, no. 3, pp. 255-265, 2003. (details)
B. Li, G. E. Christensen, G. McLennan, E. A. Hoffman, and J. M. Reinhardt. Pulmonary CT image registration and warping for tracking tissue deformation during the respiratory cycle through 3-D Consistent Image Registration. Medical Physics, vol. 35, no. 12, pp. 5575-5583, 2008. (details)
J. M. Reinhardt, K. Ding, K. Cao, G. E. Christensen, E. A. Hoffman, and S. V. Bodas. Registration-based estimates of local lung tissue expansion compared to xenon-CT measures of specific ventilation. Medical Image Analysis, vol. 12, no. 6, pp. 752-763, 2008. (details)

Related conference papers:

S. Lee, M. Abramoff, and J. M. Reinhardt. Feature-based pairwise retinal image registration by radial distortion correction. In J. P. Pluim and J. M. Reinhardt, eds., Proc. SPIE Conf. Medical Imaging, vol. 6512, San Diego, CA, 2007. (details)
S. Lee, J. M. Reinhardt, and M. Abramoff. Validation of Retinal Image Registration Algorithms by a Projective Imaging Distortion Model. In Proc IEEE Eng Med Biol Soc, vol. 1, pp. 6471-6474, 2007. (details)
S. Lee, M. Abramoff, and J. M. Reinhardt. Retinal image mosaicing using the radial distortion correction model. In J. M. Reinhardt and J. P. Pluim, eds., Proc. SPIE Conf. Medical Imaging, vol. 6914, San Diego, CA, 2008. (details)
L. Zhang and J. M. Reinhardt. 3D pulmonary CT image registration with a standard lung atlas. In C.-T. Chen and A. V. Clough, eds., Proc. SPIE Conf. Medical Imaging, vol. 3978, pp. 67-77, San Diego, CA, 2000. (details)
F. Li, C.-W. Chen, E. A. Hoffman, and J. M. Reinhardt. Evaluation and application of 3D lung warping and registration model using HRCT images. In C.-T. Chen and A. V. Clough, eds., Proc. SPIE Conf. Medical Imaging, vol. 4321, pp. 234-243, San Diego, CA, 2001. (details)

Related theses:

K. Ding. Registration-based regional lung mechanical analysis. MS thesis, The University of Iowa, Iowa City, IA, 2008. (details)
B. Li. The construction of a normative human lung atlas by inter-subject registration and warping of CT images. PhD thesis, The University of Iowa, Iowa City, IA, 2004. (details)
Y. Pan. Estimation of regional lung expansion via 3D image registration. MS thesis, The University of Iowa, Iowa City, IA, 2005. (details)
S. V. Bodas. Improved association graph matching of intra-patient airway trees. MS thesis, The University of Iowa, Iowa City, IA, 2008. (details)