A new image processing algorithm for visual prosthesis systems

Today´s rapidly developing visual prosthesis systems can provide visual perception in the form of little light spots called as “phosphenes”. Major image processing studies in this area generally use only edge information in the images to form a visual stimulation, but this approach leads to information loss in the large extent. The aim of this study, therefore, is to implement an image processing algorithm that describes the visual scene more informatively in this phosphene domain. In this algorithm, we used contrast and different patterns to describe the regions or objects with different spatial features (such as color and texture,) in the image. These different patterns allow patients to better discriminate whether the regions have similar properties or not. By this way, the visual prosthesis recipients can gather extra information of the scene, addition to the edge information and their life quality may be increased. In this proceeding, main details of the developed algortihm are explained and results are presented.