Automatic montaging of corneal sub-basal nerve images for the composition of a wide-range mosaic

Abstract

We present and discuss a computerized system able to provide a wide-range mosaic of the sub-basal nerve layer of central cornea, built from several images acquired in-vivo with confocal microscopy. The montage is performed by a fast, reliable and fully automatic computerized system that does not require any expedient or manual adjustment during the acquisition process. The resulting mosaic provides a large high quality image, which should significantly aid clinicians in evaluating and assessing in a more reliable way the pathologic signs of interest.

Automatic analysis of corneal nerves imaged using in vivo confocal microscopy

“…Despite the efficiency of the method proposed by Efron and Edwards et al, Poletti et al have argued this method ‘suffers the disadvantages of reduced montage quality with fewer sub‐basal nerve branching details’ . They used the phase correlation technique to determine the amount of offset in both translation and rotation between image pairs.…”

“…Unlike the generalised cross‐correlation method used earlier to align image pyramids in the luminance domain, the phase correlation compares information in the Fourier domain. Poletti et al used the Fast Fourier Transform to rapidly extract phase information to determine linear and angular offsets between images. Hence, phase correlation has advantages over previous methods because it directly addresses the problem of accounting for changes in the orientation of the same corneal structures seen between images due to torsional rotation of the eye.…”

In Vivo Confocal Scanning Laser Microscopy

“…For example, cSLM serves as a multifunctional tool for corneal analysis of laboratory animals [2], it allows for assessing stromal changes in patients with keratoconus before and after cross-linking [3], for experimental full-thickness corneal 3D imaging [4], and for the quantification of morphology of epithelial cell layers and the subbasal nerve plexus (SNP) [5]. Intense research is focused on large-scale image reconstruction of the SNP [6][7][8][9][10][11], because it has the potential to serve as biomarker for early neurodegenerative changes [12].…”

Automated image analysis of corneal structures in anterior-segment optical coherence tomography and in-vivo confocal microscopy images

“…Since HoG has been used originally for detection of humans in scenes it was found that coarse spatial sampling, fine orientation sampling, and strong local photometric normalization worked the best for this purpose. Poletti [79], and Allgeier [80].…”

“…Poletti et al proposed a similar algorithm based on the same distance functions [79,91] except they considered a form where the distance functions could be raised to a power, d I 1 n (i, j). In their works, the distance functions were squared while our work simply kept them as is.…”