Research

RECENT GRANTS (at Wroclaw University of Technology)

[4] OPUS-8 – Advanced imaging methods for modelling aging processes and the assessment of disease changes in glaucoma patients, ID 272166, National Science Centre Poland, 1 170 926 PLN, 11-05-2015 (current)

[3] H2020-MSCA-ITN-2014 EDEN – European Dry Eye Network, ID 642760, European Commission, 448 275 EUR, 01-03-2015 (current)

[2] FP7-PEOPLE-2013-ITN AGEYE – Aging Eye, ID MC608049, European Commission, 387 830 EUR, 01-08-2014 until 31-07-2017.

[1] Innovation Project ESP, ID 630807, Eaglet Eye, Netherlands, 60 000 EUR, 01-01-2012 until 31-12-2016.

SELECTED RESEARCH TOPICS

Eye Surface Profiler

MEASURING ANTERIOR EYE. I have been fortunate to have the  opportunity to work on the world’s first instrument for successfully measuring the topography of the entire anterior eye. The instrument, called Eye Surface Profiler and developed by a Dutch company Eaglet Eye, uses the principle of fringe projection and sophisticated image processing algorithms to measure the eye’s corneal surface, the corneo-limbal region and beyond.

It is expected that the instrument will find its applications in contact lens industry, particularly in the design and fitting of scleral contact lenses, as well as in ophthalmology where the extended anterior eye topography is of interest.


THE BOOTSTRAP. It is a well established statistical method for estimating sampling distributionsBootstrap book and SP covers based on observed data. For over a decade, together with Prof. Abdelhak M. Zoubir from Darmstadt Technical University, Darmstadt, Germany, we have been advocating this method in signal and image processing. We refer the interested readers to our tutorial paper published in the special issue of the Signal Processing Magazine (24(4):10-19, July 2007) and to our book (Cambridge University Press).

There are few reviews of the book available at:
[1] T. Wilkinson, “Book review”, IEEE Communications Engineer, 46-47, October/November 2004.

[2] G. J. Babu, “Book review”, Technometrics, 47(3):374-375, August 2005.

[3] M. Hassanali, “Book review,” Annals of Biomedical Engineering, 34(6):1074-1075, June 2006.


hsvMODELLING CORNEAL ELEVATION. Cornea is the major refracting component of the eye and contributes to 2/3 of the total optical (refractive) power of the eye, while the rest is provided by the lens. Cornea has to be aspheric to minimise the effect of spherical aberration. Knowledge of the corneal curvature is important for determining eye’s aberration, contact lens design, detecting/diagnosing corneal anomalies (e.g., keratoconus), and performing refractive surgery. Corneal surfaces can be measured using keratography, Scheimpflug imaging, or optical coherence tomography. Fitting data from such instruments with a functional parametric model leads to a more compact representation than the set of the sample points itself. In turn, such a representation is more suitable for cross-sectional analysis of corneal topography, longitudinal studies of corneal changes and studies of corneal response to accommodation, corneal classification, keratoconus detection, and even data compression.

To learn more please refer to my publications, particularly to [J6],[J9],[J18],[J23],[J40],[J46],[J52], and [J63].


eyemorphEYE MORPHOLOGY. There are many image processing techniques for extracting pupil and limbus (iris) outlines from digital images. However, the eye morphology is much more complicated and it requires assessing a large number of clinically important parameters such as upper and lower eyelid outlines, horizontal eye fissure, horizontal visible iris diameter, palpebral apertures and other parameters related to the inter-canthus line. Developing image processing techniques that would automatically extract all those parameters from digital images is challenging, particularly for down gaze and Asian eyes, whose palpebral aperture is narrower to those of Caucasian eyes.

For almost a decade I have been developing eye morphology software, i-Metrics ©, for extracting clinically important parameters from both static images and video sequences.

Some details are available in [J14],[J26] and [J28].


tfsqTEAR FILM. The cornea is covered by a thin tear film that performs various vital functions including the provision of a high quality optical surface and protecting the cornea. Traditional clinical methods for assessing tear film surface quality such as slit-lamp biomicroscope examination of corneal and conjunctival staining with vital dyes, fluorescein tear break-up time, subjective assessment of meibomian gland secretions, conjunctival hyperaemia, and evaluation of tear film volume with a Schirmer test, or cotton thread test, are mostly invasive and often unreliable. Recently, several non-invasive methods have been proposed. I have been involved in the development of three promising techniques for non-invasive assessment of tear film surface quality based on lateral shearing interferometry (in collaboration with Dr. Dorota H. Szczesna-Iskander, Institute of Physics, Wroclaw University of Technology) and high speed videokeratoscopy and wavefront sensing (in collaboration with Prof. Michael Collins, Contact Lens & Visual Optics Laboratory, QUT, Australia).

Video Examples:

  1. Evaluation of tear film surface quality on a contact lens with Dynamic-Area High Speed Videokeratoscopy  (ca. 1.8 MB, MPEG-4)
  2. Evaluation of tear film surface quality with Lateral Shearing Interferometry (ca. 2.7 MB, MPEG-4)

For more information see [J20],[J21],[J34],[J43],[J55],[J57],[J61],[J62], and [J64].


Dynamics in aberrationsDYNAMIC ABERROMETRY. Human eye is a complex but robust dynamic system in which we observe temporal changes in accommodation, temporal changes in wavefront aberrations, eye movements (micro-saccades, drift, tremor), temporal changes in surface characteristics (e.g., dynamics of tear film, corneal deformation). However, despite the observed dynamics the eye corrections are static (i.e., glasses, contact lenses, or fixed refractive surgery). Until we come up with a dynamic correction that can be applied outside a laboratory, the dynamics of eye’s optics have to be fully understood so that we could objectively arrive at the best static correction. For this, studies of the dynamics of eye’s wavefront aberrations are important. Please refer to [10],[15],[16],[27], and [29].

Wavefront aberrations are linked to the cardiopulmonary system. Recently (together with Michael Muma, Darmstadt Technical University, Germany) it was established that the considered signals are non-stationary and only through non-stationary signal analysis the complete picture of interactions between the cardiopulmonary signals and the eye’s wavefront aberration could be revealed. In fact, significant time intervals exist, where coherence between the eye’s aberrations and the cardiopulmonary signals is high. More information can be found in [J59].

Wavefront dynamics


OCULAR PULSE. Measuring ocular pulse is important in a range of medical applications where the assessment of hemodynamic status of the eye is required. For example, low amplitudes of the ocular pulse could indicate glaucoma and several other eye diseases which may have vascular aetiologies. The delays in the ocular pulse, on the other hand, could indicate the presence of severe carotid artery occlusive disease.

Taking the simplest approximation of the eye globe as a sphere where the volume is proportional to the radius cubed while changes in the volume are proportional to changes in radius, we (together with Prof. Henry Kasprzak and Dr. Malogrzata Kowalska, Institute of Physics, Wroclaw University of Technology) postulate that there must be a relationship between the ocular pulse and the speed of the observed radial displacements of the corneal apex similarly to that found between ocular pulse pressures and retinal vessel velocities. That is why we aimed at studding the characteristics of the longitudinal eye movements (LEM).

More information can be found in [J29],[J39],[J44],[J49] and [J66].

Eye movements and ECG


OBJECTIVE REFRACTION. With the increased popularity of wavefront sensor devices there hasrefzer been significant interest in relating the subjectively measured sphero-cylindrical refractive error to the objectively measured wavefront aberrations. The standard way (yes, there is already an ANSI standard there) is to use a set of estimated Zernike polynomial coefficients (either the second order or the higher order representation). I have been working on this problem for some time first with Brett Davis (School of Optometry, QUT, Australia) and we ended-up deriving a better representation that we called refractive Zernike power polynomials [J35]. This work sparked some interest that resulted in a new set of functions (derived by Jayoung Nam, Indiana University) that are called Zernike slope polynomials [J48],[J50],and [J51]. These new representations describe aberrations in dioptres and are more intuitive to optometrists and ophthalmologists.

We have been awarded the J. Lloyd Hewett Award for Excellence for 2010 (for the best paper in years 2007-2009) for the work published in [J51].