A Study on Significance of Color in Face Recognition using Several Eigenface Algorithms

Face recognition has rapidly evolved and has become very popular in recent years. It is being intensively researched and found many applications, primarily in biometric security systems. One of the main challenges in face recognition is to identify different features playing fundamental role in face description. The role of color, which appears to be a salient attribute of faces, is still debatable within the literature. Some research has suggested that it confers little recognition advantage for identifying faces while other research suggests that color is a very important cue. The clear perception of colors in the environment illustrates that color must be important in the interpretation of complex scenes and recognizing objects in the environment. In this paper several face recognition algorithms are reviewed compared using both grayscale and color images. The accuracy of each algorithm has been determined and they were ranked according to recognition rates. A system is also proposed, which uses color features for face recognition. This system can be used by different face recognition algorithms. The experiments have been performed on two wimagesell-known databases: CVL database containing 114 individuals with 7 images per individual and Georgia face database consisting of 50 individuals with 15 pictures of each individual taken with different poses, expressions and backgrounds.