A FAST, ROBUST PATTERN-RECOGNITION SYSTEM FOR LOW-LIGHT-LEVEL IMAGE REGISTRATION AND ITS APPLICATION TO RETINAL IMAGING

We describe an image processing system which we have developed to alig n autofluorescence and high-magnification images taken with a laser sc anning ophthalmoscope. The low signal to noise ratio of these images m akes pattern recognition a non-trivial task. However, once n images ar e aligned and averaged, the noise levels drop by a factor of root n an d the image quality is improved. We include examples of autofluorescen ce images and images of the cone photoreceptor mosaic obtained using t his system. (C) 1998 Optical Society of America.