[An adaptive method for image enhancement of retinal vascular images based on morphology and human visual properties]

A method of contrast enhancement for retinal vascular images is provided. Opening of a gray-scale image by a gray-scale structuring element(SE) can generate a background image. Enhancement is achieved by subtracting this background image from the original image and then applying a adaptive gray-scale modification technique which is based on the local contrast of the image and the lobserver's Just-Noticeable-Difference(JND). As the method includes certain human visual properties and treates smooth and detail areas of a retinal image differently, the common problems of producing ringing artifacts around sharp transition regions and excessive enhancement of noise for many traditional enhancement methods are avoided. Another advantage of this method is the retention of vessel dimension which directly benefit from using morphological opening filter. The suggested method performed very well both analytically and visually.