Purchase this article with an account.
A.B. Hansen; Image Quality of Digital Non-mydriatic Retinal Images . Invest. Ophthalmol. Vis. Sci. 2003;44(13):3651.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To set criterias for the evaluation of image quality of digital non-mydriatic retinal images, and evaluate photographs obtained with and without mydriasis using the same digital camera. Methods: 83 patients with either type 1 or type 2 diabetes were recruited at The Steno Diabetes Center based on medical records. All patients were photographed with and without mydriasis using a TOPCON TRC-NW6S non-mydriatic digital fundus camera interfaced with a JVC 3CCD color camera with a pixel resolution of 1450 x 1062. Five overlapping 45° images (posterior pole, nasal, temporal, superior and inferior) were captured of each eye, and using an IMAGEnet 2000 computer system the images were merged into a single mosaic image and stored in TIF format. Image quality was then evaluated by two ophthalmologist based on the criterias listed in tabel 1. Results: The quality of all images captured with mydriasis were evaluated as being either excellent or good (tabel 1), blur due to cataract being the primary reason why not all mydriasis images were evaluated as being excellent. The quality of the images captured without pupildilation were evaluated as listed in tabel 1. As the non-mydriatic camera has as a requirement, that the pupil diameter is to be greather then 4 mm to obtain an adequate image quality, it was not surprising that the primary reason for undilated images to be evaluated as being either poor or ungradable were due to the pupildiameter being < 4 mm. This was shown to be highly age related (P<0.001 Χ2trend). Thus non among the 20-29 year old had a pupil diameter < 4 mm, 7 % of the 30-39 year old, 12 % of the 40-49 year old, 16 % of the 50-59, 62 % of 60-69 year old and 100 % of the 70-79 year old had a pupil diameter < 4 mm. Cataract did not influence the evaluation of undilated images substantially. Conclusion: When using a nonmydriatic digital camera in screening for diabetic retinopaty approximately 1 out of 5 will require mydriasis. Table 1. Criterias and distribution of image quality View OriginalDownload SlideView OriginalDownload Slide
This PDF is available to Subscribers Only