Abstract
Abstract: :
Purpose: Eccentric photorefraction has been used to screen for refractive error, however, conventional cameras with telephoto lens and film are costly and involve a delay to view results. Digital cameras available for family use can have integrated flash and zoom control, and offer instant viewing of captured images plus reduced costs of consumables. The usefulness of a digital camera for photorefraction was evaluated using both a model eye and comparison of subjective prescriptions with photorefractions of young adults. Method: Flash still photos of a model eye at various dioptric settings were taken at different working distances with a JVC GR-DVM70 DualCam saving to miniDV tape. Pupil crescent width was measured on downloaded images (NIH freeware Image) and refraction computed via revised formulae that excluded the camera aperture from calculations, as some specifications for the camera were unavailable. 52 students aged 17-30 years were photographed using the optimum distance derived from the model eye investigation. One photo was taken and immediately classified off the LCD screen as myopic or not. If not, another photo was taken through a pair of +1.50 DS, which simplified the interpretation of any photos presenting little crescent, and permitted classification based upon screening pass criteria of -1.00D to +0.50D. Refractive information derived from the photos was compared with subjective refractions held on clinic records. Results: Using dioptric settings of the model eye between -5.00D to +5.00D and working distances of 1.5m to 2.5m, powers estimated from the image crescent width agreed at higher powers of both myopia and hypermetropia, but underestimated the magnitude of myopia (by up to 0.75D) and hypermetropia (by 0.25D) at lower powers. Comparison of photorefraction and subjective refractions indicate a sensitivity of 95% and specificity of 80% for photorefraction using this camera. All eyes were clinically classified appropriately (i.e.within±0.25 D). Image analysis has now been semi-automated with the writing of a small program to permit 'drag-and-drop' circles to compute pupil and reflex sizes and allow calculation of the refractive error directly. Conclusion: Digital still flash photography at video resolution can match existing conventional cameras, but offers advantages of practicality. For screening purposes, results with an off-the-shelf digital camera were comparable with other dedicated commercial digital photoscreeners, but may represent better value.
Keywords: 542 refraction • 430 imaging/image analysis: clinical