Perceptual-based Correction of Photo Red-Eye

R. Ulichney and M. Gaubatz (USA)


Digital photography, image enhancement, red-eye


Many recent works in digital photograph image enhancement have included methods for detecting camera-flash-induced red-eye artifacts. Less effort, however, has been spent on the correction procedure itself. A variety of corrective procedures have been presented that reduce the chrominance and luminance of detected pixels in different ways. Given that the original pupil color of a subject is often unrecoverable, a simple chrominance de-saturation effectively removes the red hue from the artifact pixels. This work presents the results of a perceptual study designed to find the most visually pleasing target luminance for corrected images. The results of this experiment are analyzed and correlated with the test data to yield simple equations for target luminance. These are in turn combined with a fully automated procedure designed to minimize intrusive effects associated with pixel re-coloration. Data and equations are presented for hard-copy (printed) and soft copy (on-screen) images.

Important Links:

Go Back