Superpixel based retinal area detection in SLO images

Distinguishing true retinal area from artefacts in SLO images is a challenging task, which is the first important step towards computeraided disease diagnosis. In this paper, we have developed a new method based on superpixel feature analysis and classification approaches for determination of retinal area scanned by Scanning Laser Ophthalmoscope(SLO). Our prototype has achieved the accuracy of 90% on healthy as well as diseased retinal images. To the best of our knowledge, this is the first work on retinal area detection in SLO images.