Retinal Image Enhancement based on the Human Visual System

DESCRIPTION (provided by applicant): The objective of this project is to demonstrate a methodology for improving the quality of retinal images taken with standard fundus camera. Kestrel has developed a family of algorithms that perform image retinal enhancement through the implementation of a methods that are inspired from human visual system mechanisms. Standard retinal images often suffer from illumination artifacts, large dynamic range in the reflectance, and blurring. In this project Kestrel will demonstrate its image quality enhancement algorithms on a wide variety of fundus camera systems, including a digital non-mydriatic device, a mydriatic film camera, and a mydriatic mega- pixel instrument. Our objective is to show, through controlled evaluations by ophthalmologists and highly trained ophtalmic technicians, that our processed images result in significant improvements in image quality and which in turn would result in improved sensitivity and specificity for grading or screening retinal images for diseases such as diabetic retinopathy and age-related macular degeneration. The commercial potential of this software is immense. A low-cost means for improving image quality will be of interest to all retinal camera manufactures. The project will show that improvement can be realized for any existing fundus camera, thereby opening up the market to all future and existing instruments.