A spectral-discrimination method for tear-film lipid-layer thickness estimation from fringe pattern images.

Abstract

Examination of the tear-film lipid layer is often helpful in the prognosis of prospective contact lens patients and contact lens related problems, and in the analysis of symptomatic noncontact lens-wearing patients. In particular, the thickness of the lipid layer is considered to be an informative cue in studying the tear-film stability and uncovering of certain disorders. We propose a method for the accurate estimation of the lipid-layer thickness, exploiting the intensity and color information in Fizeau fringe images. The technique is based on a quantitative measure for discriminating among the spectra associated with different thicknesses. We propose an optical system for imaging the interference patterns, develop a mathematical model based on the physics of the fringe formation and sensing, and describe the calibration of the optical system using this model. The thickness extraction is readily carried out utilizing a lookup table. The proposed method would enable objective evaluation of the lipid layer characteristics, and provide a means for examining the dynamic changes in its thickness and spatial distribution during inter-blink periods.