Figure 3From: Human recognition based on retinal images and using new similarity function Autocorrelation principal curvature space. Heavy lines give corner, edge, and flat classification, and fine lines are equi-response contours [21].Back to article page