Purpose
To demonstrate a smart, low-cost, user friendly non-mydriatic retinal camera for use in underserved rural and urban communities.
Methods
Portable, low-cost cameras have been introduced to the commercial market; however, at $8,000 to $12,000 these cameras remain unaffordable for many clinics. Smartphone cameras have been introduced to leverage their popularity and portability but produce inconsistent image quality. These cameras will be most useful where there is no staff with experience in retinal imaging. Therefore, the need is not only for a low-cost camera but one that is also easy to use and consistently produces high quality images without mydriasis when operated by minimally trained personnel. This study presents the Smart i-Rx, a low-cost camera that meets these requirements (FIG. 1). Consistent image quality is achieved through real time feedback to the user via advanced digital image processing. The Smart i-Rx was designed using off-the-shelf components that in volume cost less than $1,000. Tests with minimally trained individuals were conducted and images compared to a commercial Canon CR1 non-mydriatic camera. The camera was tested on 20 controls and 20 diabetic subjects. Qualitative image quality was assessed by a certified ophthalmic technician using a 5-point Likert scale. Quantitative image quality was assessed by a computer algorithm that uses both objective and subjective features (FIG. 2).
Results
Qualitative and quantitative image quality were not significantly different between the cameras (p<0.05). There were 10 cases that presented retinal abnormalities. Expert assessment of retinal abnormalities was not significantly different between the cameras (p<0.05).
Conclusions
This study shows that a low-cost, non-mydriatic camera which produces high quality images is feasible. The introduction of such a camera to clinics in underserved and economically depressed communities would reduce significantly the cost of screening large populations. The integration of real-time automatic image quality assessment would reduce the need for extensive training for use by minimally trained staff, thus reducing the two main factors affecting the cost of screening, i.e. equipment and staff.
Keywords: 550 imaging/image analysis: clinical •
688 retina •
499 diabetic retinopathy