Introduction to Face Recognition and Recent Work

Tianrui Liu,a,b Wan-Chi Siu,a Cigdem Turan,a Shun-Cheung Lai,aPerhaps, face is one of the most useful biometrics for person recog-nition. Conventionally, three steps are required for face recognition,namely face detection, face alignment and face recognition. In thischapter, we just briefly mention the first two steps and concentrateon some core technologies for face recognition. We will start thechapter with a thorough introduction to using Eigenface or PrincipalComponent Analysis (PCA) for face recognition, since it is a classicapproach, and many modern approaches take it as a reference fordevelopment. Subsequently, we briefly describe Linear DiscriminantAnalysis (LDA) and other modern approaches, including subspacelearning methods and deep learning neural networks for facerecognition. We expect that the beginning part of this chapter willbe useful for beginners, whilst the rest of the chapter could form agood reference for early researchers in the area.