A method of automatic red-eye removal
Pengfei Shi · Journal of Circuits and Systems · 2006
This paper presents an automatic red-eye removal method based on face detection in digital photos.The method can be described as follows.First,face regions are detected through the object detection method based on Haar-like features and the AdaBoost learning algorithm.Then the detected regions are classified as typical red-eye or weak red-eye by the redness distribution in the regions and the red-eye mask is formed accordingly.Finally,the color correction and edge smoothing are applied to the red-eye regions.The experiments prove the efficiency of this method in removing the red-eye effect.The results also show that it is robust with fast processing speed and adaptable to embedded applications.