Title :
Automatic red-eye detection and removal
Author :
Miao, Xiao-Ping ; Sim, Terence
Author_Institution :
Sch. of Comput., Nat. Univ. of Singapore
Abstract :
Red-eye is a very common problem in flash photography which can ruin a good photo by introducing color aberration into the subject´s eyes. Previous methods to deal with this problem include special speedlight apparatus or flash mode that can reduce the red-eye effect, as well as post-capture red-eye correction software. The paper presents a new approach to detecting and correcting red-eye defects automatically by combining flash and non-flash digital images. It is suitable to be incorporated into compact digital cameras that support continuous shooting. Such a camera would eliminate red-eye immediately after image capture. Unlike existing approaches, our method is simple, fast and can recover the true color of the eyes
Keywords :
image colour analysis; image restoration; automatic red-eye detection; automatic red-eye removal; color aberration; compact digital cameras; continuous shooting; flash digital images; flash photography; nonflash digital images; red-eye correction; Biomedical imaging; Blood vessels; Contracts; Detectors; Digital cameras; Eyes; Face detection; Humans; Iris; Photography;
Conference_Titel :
Multimedia and Expo, 2004. ICME '04. 2004 IEEE International Conference on
Conference_Location :
Taipei
Print_ISBN :
0-7803-8603-5
DOI :
10.1109/ICME.2004.1394434