mobileVision: Towards a patient-operable, at-home, non-mydriatic retinal imaging system

Abstract

We present our work towards a patient-operable, at-home, non-mydriatic retinal imaging system. Early detection of ocular diseases like diabetic retinopathy is facilitated by retinal examinations that typically include imaging conducted by experienced examiners using sophisticated and expensive equipment. This creates a barrier to care for patients, preventing retinal examinations and leading to worse patient outcomes. To address this problem, we propose a system operated directly by the patient; a specialist need not be present. An adjustable fixation target, indirect imaging optics, and a self-contained illumination system comprise a robust yet compact attachment which affixes to the patient's smartphone. This allows the smartphone's camera to acquire digital retinal images (1MP, 23 degree, 10um resolution) under controlled eye rotation while preserving eye-camera pupil matching through mechanically constrained translation of the optics. A software application downloaded to the smartphone assists the user in imaging pre-set retinal fields: it automatically controls the fixation target and illumination while giving verbal instructions. The software combines the resulting image sequence to form a 60 degree retinal panorama which can be sent electronically to a health care provider for remote viewing. We have constructed and used a preliminary prototype to image rat and human eyes in vivo. Our aim is to enable untrained persons to image their own retinas and those of others, thereby reducing burdensome demand for both health care providers and costly imaging instrumentation. This device has the potential to become a valuable tool for self-monitoring in patients requiring frequent retinal examinations.