Estimating Corneal Surface Topography in Videokeratoscopy in the Presence of Strong Signal Interference

Videokeratoscopy techniques rely on a number of factors in order to achieve accurate estimates of corneal surface topography. Good tear film quality, minimal reflections from eyelashes, and minimal eye movements are essential for corneal topography estimates to be reliable. However, in practice, these ideal conditions may not always be fulfilled, especially in cases of subjects diagnosed with dry eye syndrome, having narrow palpebral apertures, long eyelashes, or nystagmus (uncontrolled eye movements). Such nonoptimal conditions of image acquisition result in poorer estimates of corneal topography. The aim of this paper was to devise a technique that would provide more accurate estimation of corneal topography in such situations and particularly when the source of signal interference is strong. This was achieved by developing a set of algorithms that extract the interference from the acquired raw videokeratoscopic image and filter the topography according to the interference location. The experiments carried out with test surfaces and real corneas showed that this new technique leads to a significant improvement in the topography estimator. Additionally, it is an interference indication procedure that, in the future, could be used for the purpose of tear film quality estimation.