Premium
Fully automated corneal endothelial morphometry of a large set of images captured by clinical specular microscopy
Author(s) -
BUCHT C,
SÖDERBERG PG
Publication year - 2010
Publication title -
acta ophthalmologica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.534
H-Index - 87
eISSN - 1755-3768
pISSN - 1755-375X
DOI - 10.1111/j.1755-3768.2010.4334.x
Subject(s) - corneal endothelium , software , fourier analysis , computer science , specular reflection , fourier transform , artificial intelligence , image processing , computer vision , image analysis , biomedical engineering , cornea , digital image processing , optics , image (mathematics) , mathematics , medicine , physics , mathematical analysis , programming language
Purpose The endothelial cell density is the most important morphological factor of the corneal endothelium. Morphometry of the corneal endothelium is an important part of several clinical applications. Morphometry of the endothelium is presently carried out by semi automated analysis of pictures captured by Clinical Specular Microscopy (CSM). The need of operator involvement makes this process time consuming. This study presents a method for fully automated analysis of a large range of in vivo images of the corneal endothelium, captured by CSM, using Fourier analysis. Methods Software was developed in the mathematical programming language MATLAB. Pictures of the corneal endothelium, captured by CSM, were read into the analysis software. The software performed automated digital enhancement of the images. The enhanced images were Fourier transformed, using the Fast Fourier Transform. Relevant characteristics of the Fourier transformed images were identified and sampled. The data obtained from each transformed image was used to calculate the mean cell density of the original image, which in turn was compared to a semi automated method cell density estimate. The calculation was based on well known diffraction theory. Results Estimated cell densities of the corneal endothelium were obtained, using fully automated analysis software on 292 images captured by CSM. Using linear regression, a relatively large correlation between the estimates of the fully automated method and the semi automated method was found. Conclusion The results using the considerably faster fully automated method are highly encouraging for further development and implementation of the method.