Development and testing of a stabilization and image processing system for improvement of mobile fundus camera image quality

The field of teleophthalmology has been expanding in recent years due to the development and advancement of mobile fundus cameras. Many of these devices involve taking a video of the retina using a smartphone coupled with an attachable lens system. Despite recent advances, the videos that are obtained with these devices can sometimes be difficult to use for clinical purposes. This is because the videos often have a small field of view and can include blurry frames and blinks. In order to improve the ease at which these videos can be obtained and interpreted, a stabilization system and Android application for processing these videos was created. This Android application is intended to exclude unusable video frames, and stitch together the remaining frames into a single image. The effectiveness of this application is tested using videos obtained with the D-EYE device. These images are then compared to clinical images, and images obtained from the DEYE without the image stitching application. In this study, the image quality of the stitched images was found to be worse than that obtained from a clinical device.