Automatic Redeye Correction in Digital Photos

A methods to correct the artifact known as “red eye” is proposed by means of digital color image processing and classification procedure. First, skin like regions is detected with a pixel-based support vector machine processing; morphological operations are then used to eliminate the extra areas. In the second step 6 new features include geometric and color metrics are proposed to better classification of artifacts. Finally a support vector machine is used to classify the output of skin detected images by the use of presented features. 30 custom images are used to accuracy evaluation and results show the high performance of the proposed method toward some other methods. Keywords Redeye correction, HSI color space, Skin detection, Classification, Support Vector Machines, Morphological operations, Color and Geometry features 1. INTRODUCTION effect The red eye effect has been a major problem in photographers since they first started including flash guns in their cameras; it happens when using a photographic flash very close to the camera lens (as with most compact cameras); the extent use of compact devices with built-in flash, involving cell phones and handheld computers, causes a large amount of digital photographs that are possibly affected with red eye. As shown in fig.1, an extreme flash enters the pupil from a direct angle, mirrors the blood vessels in the back of the eye, and leaves the pupil headed straight back to the camera. The light picks up most of its red color from blood in the choroid, a vascular part of the eye which supplies blood to the retina causes the exiting light red. To decrease these artifacts, most cameras have a red-eye flash mode which fires a series of pre-flashes prior to picture capturing [1]; despite detection and correction of red eye is a regular characteristic on most digital cameras today, the technology is not one-size-fits-all. Its efficiency changes based on many causes, including lighting conditions, topic and camera angle. Understanding what causes red eye effect is a good first step toward building a better camera for the makers. Fast pre-flashes make pupil contraction, thus, minimizing the area of reflection; such operation does not entirely remove the red-eye , but reduces it. Power consumption is the major drawback of the pre-flash approach in order to the most power consuming of flash. Besides, repeated flashes usually cause uncomfortable feeling. Various attempts have been made to reduce this problem at the time the photograph is taken.